Let's discuss why AMD is no longer an enthusiast platform

[CleanSlate]

With the last few product roll outs and the recent quote from AMD's CEO we've seen AMD shift into this quasi-useful-processor creating company and I'm personally mad as a bumblebee shaken up in a plastic bag over this but I'm curious what the community has to say about it.

Is AMD going to survive into your next rig while putting out these processors that are so under-performing that Intel basically has no competition at any where near the high end?

Is this the first step towards the death-knell for AMD? We've seen AMD battered again and again attempting to take the performance crown from Intel but never quite being able to be dominant in anything but the price-to-performance ratio for under-performing chips in certain categories of performance.

AMD: why allow a monopoly on performance by your direct competitor in all categories while fighting for market share in a market that takes the performance crown of each generation extremely seriously?

 

[Tsumi]

AMD never said that they were quitting out of the enthusiast sector. They just said they will no longer focus on it as their primary source of income, and will instead focus on mobile and increasing efficiency for servers, which makes sense because those two markets are several magnitudes bigger than the enthusiast market. They will scale up the processors as much as possible for their top end enthusiast CPUs, but that's not the priority.

Threads like these are pretty useless because they're completely one-sided, and acts as if the enthusiast market was everything. Mobile, OEMs, servers, and businesses all prioritize efficiency over raw power.

Also, don't act as if AMD isn't trying. I'm pretty sure they're trying as hard as they can, but more than 3 years ago they bet on a future that simply did not happen. They bet on a future where multitasking and multithreaded workloads would be a lot more commonplace, and it didn't happen. Their funding is also significantly less than Intel.

 

[stealthy123]

I think the most profitable markets aren't in bleeding edge performance.

AMD could execute some solid decisions, become profitable again, and then shock the world.

But, capitalism being what it is, they are beholden to the share holders, and pretty much no one else. They need to make the shareholders money, and once that starts happening, they may have the freedom to make something innovative.

 

[DedEmbryonicCe11]

AMD: why allow a monopoly on performance by your direct competitor in all categories while fighting for market share in a market that takes the performance crown of each generation extremely seriously?

What? Do you honestly think AMD has the resources to compete with Intel >$200 in the single socket desktop market? Apparently that's what you consider "performance" to be since the 8350 is very well positioned at $195 MSRP when you consider all the applications out there that aren't games. You would do yourself a world of favor if you simply realized that short of every engineer at Intel suffering a traumatic brain injury over the next few years there isn't every going to be another top of the line desktop processor from AMD that trounces Intel's offerings. There's obviously a crowd (myself included) that would have killed for a Phenom III with Llano improvements on 32nm instead of the BD architecture. If you look at the recent round of benchmarks though, it's clear that even that would have lost, and lost badly in many scenarios to Intel. It was lose/lose and AMD went with a design that made more sense for servers and business applications.

 

[CleanSlate]

What? Do you honestly think AMD has the resources to compete with Intel >$200 in the single socket desktop market? Apparently that's what you consider "performance" to be since the 8350 is very well positioned at $195 MSRP ...

The whole point of a business is to invest in itself and create something that makes the company money. In this case it's AMD investing in engineering a processor that can compete with Intel in the >$200 market.

That's why they've done it for years and that's why they should be doing it now.

As well positioned one processor is or not, AMD needs the performance crown to survive in the long haul. Period. Otherwise it will go the way of the dodo.

 

[Tsumi]

The whole point of a business is to invest in itself and create something that makes the company money. In this case it's AMD investing in engineering a processor that can compete with Intel in the >$200 market.

That's why they've done it for years and that's why they should be doing it now.

As well positioned one processor is or not, AMD needs the performance crown to survive in the long haul. Period. Otherwise it will go the way of the dodo.

If they can have a $200+ processor for the enthusiast market, they would. But they can't, so deal with it. And as I've said, raw performance is nowhere near as important as efficiency, enthusiasts are a minor part of the market.

And if you want to look at $200+ AMD processors, just look at Opterons. AMD has expensive processors, just none that would compete in the enthusiast sector.

 

[octoberasian]

I've probably said it many times over and over again, and many different ways already on these forums. Here's one such post:

http://hardforum.com/showpost.php?p=1039245098&postcount=28

You have to think realistically here, especially from a business standpoint.

Yes, I agree that any and every company needs a product to sell in order for the company to make money. However, it takes money to make money. To get a better product out for the market means understanding market trends and where the money is, and take advantage of it. In effect, you make a product that will sell and will sell well if you understand the current market conditions.

The current market is shifting towards more mobile products and more energy efficient products, especially in the automotive market. Toyota Prius, for example, is the number one vehicle sold in California because our gas prices here are expensive. Car shoppers here are buying more hybrid vehicles and vehicles with high fuel economy ratings.

But, back to AMD. AMD has neither the cash flow nor the revenue to invest more in a higher performing CPU that'll match Intel. Not now and not ever. They have to take their time to make a good product, like any company does on a low budget. What may take Intel a year or two to do in their tick-tock cycle, may take five years for AMD on a smaller budget.

That's the reality of it.

We won't see any significant improvement in performance until Steamroller or Excavator coming out of AMD. AMD is putting its money where it believes and sees where the market will eventually head to. They've invested heavily in HSA (GPU and CPU as a cohesive compute platform), purchase Sea Micro server company, investing into cloud computing technologies, and directing their efforts into mobile and notebook APUs. Why? The market is going that way and that's where the money is. They've also focused their attention on multi-core performance because in the server space, that matters. In the future when software catches up to hardware, they, AMD, believe it'll be a multi-threaded software market.

To compete on the level of Intel is going to take significant amounts of investment in a market that's steadily declining and slowing year-over-year-- desktop processor market. Not only is it because of a sluggish economy, but the market is responding with their wallets with smaller computing devices. Intel knows this and have admitted it themselves when their revenue from this past fiscal year was less than last year's. They may not have given up on the desktop market, and neither has AMD because there is still some money to be made in the desktop processor market. That's even if the profits or ROI (return on investment) is much smaller than the mobile processor market.

ARM themselves have reported great revenue this fiscal year. Qualcomm as well, Apple also. They're making money is what's soon to be a very crowded mobile market.

You have to be realistic. As a business if you have only X-amount of money to invest in, you can only stretch that dollar so much that certain things will get more focus than others. In this case, for AMD, that's the desktop processor market. Have they given up on it? No. Have they quit the market entirely? No. Have they shifted their focus like any business should? Yes.

I have said this in another post:

Ok, this is going to be a stretch but looking over at various places and AMD's slides, my best estimates are the following:

• First-half 2013 - "Steamroller"-based APU on 28 nm process (Kaveri)
  • Integrated FCH (Fusion Control Hub); yes, on-die instead of a separate external chip.
  • GCN core-based on-die GPU. older 5000- and 6000-series cores no longer used
  • First possible use of HSA (hetergenous system architecture) with GPU as a co-processor or compute core.
  • Low power and ultra low power variants Kabini and Temash will be released sometime in 2013 alongside Sea Islands (Radeon 8000-series)
  • microOP-like cache
  • Dual 4-wide decoders per integer core
  • Improved frontend
  • Improved L1 cache
  • Dynamic L2 cache
  • Improved power consumption
• First-half 2014 - "Steamroller"-based FX and Opeteron processors on 28 nm process
  • "Greater parallelism"
  • Socket AM3+
  • microOP-like cache
  • Dual 4-wide decoders per integer core
  • Improved frontend
  • Improved L1 cache
  • Dynamic L2 cache
  • Improved power consumption
• Second-half 2014 - Possible release of "Excavator"-based APU
  • High-density library first used
  • Greatly improved power consumption
  • Second possible use of HSA and probably first AMD processor to fuse both GPU and CPU cores as one component
  • Possible use of Sea Islands GCN cores for on-die GPU
• First-half 2015 - Possible release of "Excavator"-based FX and Opteron processors
  • Greater compute capabilities
  • Greater cohesion and communication between external GPU and CPU
  • Third possible use of HSA and on-die GPU and CPU core cohesion.

It honestly looks like a quasi-"tick-tock" cycle for AMD.

Current unknowns for Excavator:

• 20nm process?
• 3D transistors?
• New architecture? First realization and first use of AMD's "fusion" of GPU and CPU cores?
• Cache size and improvements including power consumption?
• Socket and pin-count? (Hint: It's no longer AMx-based socket)
• On-die northbridge (PCI-E 3.0?) and/or on-die southbridge? (High-density library hints that more can be squeezed in a much smaller space.)
• HyperTransport completely gone in a the desktop/mobile (non-server) space? (Note: HyperTransport is still a part of the AMx-socket structure and external I/O communication. Replace AMx socket, you no longer have HyperTransport and, in effect, no backwards compatible with AM3+, so something has to replace it...)

Sources:
http://en.wikipedia.org/wiki/Bulldozer_(microarchitecture)

http://ir.amd.com/phoenix.zhtml?c=74093&p=irol-2012analystday

http://www.anandtech.com/show/6201/amd-details-its-3rd-gen-steamroller-architecture

If AMD was quitting the market entirely, they would not have planned that many improvements to the original Bulldozer architecture. It's entirely possible now that the Excavator module will look nothing like Bulldozer before it. AMD is moving towards a multi-threaded future, even as foolish it may seem. And, they're moving their processors towards a cohesion of the GPU and CPU as one unit. The smaller, single and shared floating point unit in each module may be indicative of that direction-- moving the FPU computational work onto the more mathematically powerful GPU cores. We may not see it now or immediately, but servers will probably be the first to benefit from it. Consumers will probably see this benefit by Steamroller or by Excavator. However, that's a few years from now. By the time these are out, Intel will have released Broadwell and its successor.

Why is AMD behind? Well, in part, blame a lot of the issues at their management and executive levels. AMD quitting? No. Laying off 20% to 25% of your workforce isn't quitting the market. Until they start filing for Chapter 11 bankruptcy, that is when we'll start worrying. Every company, especially in this current economy, are facing tough decisions to stay afloat and ensure they still have enough operating income to run the company. It's restructuring the company to make sure there's still enough money to invest in current and future products, and to keep the company running. What AMD is doing is no different than no other company that has done the same since the financial crises started some years ago.

AMD is taking risks, and they're betting on it with the company's coffers. They're hoping it'll pay off in the end.

Above all, you have to realize AMD is nowhere near the level of Intel. If they were and their processors were as fast as Intel's, we'd probably see a higher revenue each fiscal year and possibly processor pricing near or equal to Intel's.

IF you think you can turn the next AMD processor to match Ivy Bridge or Haswell with an average revenue per fiscal quarter of $1.2 to $1.5 billion dollars, then go right ahead and call AMD: http://www.amd.com/us/aboutamd/contact-us/Pages/contact-us.aspx

IF you think have the talent to miraculously turn the next AMD processor into the next big thing that kicks Intel's ass, go here:
http://www.amd.com/us/aboutamd/careers/Pages/careers.aspx

I'm sure AMD will need talented smartasses to help turn their company around. They probably need more talented engineers and chip designers than smartasses though.

Intel can afford both, of course. Heck, if they wanted to, they'd probably have a controlling share of Microsoft because they could probably afford it. (Of course, I'm kidding. Microsoft would probably do that instead.)

Everytime I see these kinds of threads of AMD, I think of the following:

• People have not taken business or economic courses in college.
• People are overly pessimistic, aside from realistic.
• People are not thinking very smartly on how to work on a shoestring budget.

I commented before in another thread:

It's like AMD telling its people they have $500 to last for the month and come up with something amazing. However, Intel would likely be: "Here's $5000 for the month, do something with it."​

Or, another way:

For AMD, it'd probably be taking something like $5,000,000 and slowly invest it over 5 years for processor improvements or new processor technologies. For example, we won't see the results of AMD's HSA investments until maybe Excavator, or Steamroller at the earliest.

Whereas for Intel, it'd probably be taking ten times that-- $50,000,000-- and invest it every two years for each processor tick-tock cycle to improve things like execution units, cache size, power consumption.

And, Intel could add another $50,000,000 to invest in a new architecture and/or materials over five years. For example, 3D transistor technology wasn't conceived by Intel until the early 2000s with products first implementing it in Ivy Bridge in 2012.​

That's the brutal reality of it, and you and everyone else will have to get that through their thickheaded skulls to realize it.

AMD IS NOT INTEL AND NEVER WILL BE UNLESS WE START BUYING THEIR PRODUCTS AND/OR HELP INVEST IN THE COMPANY.

Am I an AMD sympathizer? Yes and no.

Yes, because innovation relies on competition. The next "big thing" only exists because someone else tried to do it better. Without competition, the market stagnates. It's simple economics, and everyone learns this at one point of their educating life. If AMD were to cease to exist by next year or two years from now, Intel will stagnate in the long run. We may not see the effects immediately, but it will be seen. Instead of Haswell by next year, we'd probably see it in 2014 or 2015 because why release a new processor when the current one is still make money? Or, why release a new processor when your current product isn't even being challenged? Eventually in such a case, we'd probably see prices slowly rise in the near future because competition no longer exists. That is why we need AMD in the processor market, even if they've shifted their focus towards a low power and mobile processor market.

No, because AMD has had pretty damn shoddy executives making piss-poor decisions and probably started with Ruiz. Read is going to try to do something about it seeing how well he's done at IBM (or was it Lenovo?) before it. However, he's working with rather incompetent people at the executive and management level that needs to be sorely addressed now instead of later. They're also taking huge risks in an unforeseen market-- multithreaded software. Software developers have yet to move beyond single threaded applications in the majority of software today. Most use two threads, but never more than four. The only market that benefits from it the most is the server market-- scientific, government and enterprise customers. AMD is shifting its processors steadily towards greater parallelism (Steamroller) to greater compute capability (Excavator and beyond). But, are the risks worth taking? Every business takes risk, but it has to be risks worth taking that will give you a better ROI. I, myself, is both worried and optimistic at the risks AMD is taking. They've invested so much on a shoestring income that they have spread themselves too thin. We'll have to wait and see if any of those investments pay off in the end. No one can predict the future, but in economics, consumers-- whether private, educational, or government-- decide where the market goes. At the moment, consumers like you and I are telling the market we'd like more mobile products. So, we'll have to see if this pays off for AMD.

In the end, and in my honest opinion, if AMD wants to survive, license ARM architecture for your APUs. Make both x86 and ARM APUs combined with your Radeon cores. Attack the market from both sides. ARM is already a customer at GlobalFoundries and will be the first to see 3D transistors from GF. ARM is already a member of the HSA Foundation for greater cohesion of CPU and GPU for GPGPU compute. ARM is also already a partner with AMD to make security cores for AMD's mobile products. Why not take it further? A mobile AMD ARM processor with Radeon GCN cores may shake the market up, and push Intel and Nvidia to do better in that space, along with Qualcomm.

Be realistic here. We cannot expect nor should we ever expect AMD to be Intel. Their revenues and operating incomes are at around an 8-to-1 or 10-to-1 ratio favoring Intel. That is the harsh reality of it. AMD would have to give up their current investments to make a better single-threaded performing processor to best Intel.

But, is it smart? No. The high-end desktop processor market will not give AMD enough money to survive on. Only Intel can afford those risks and losses in the slowly declining desktop processor market, not AMD. And, only Intel can afford to improve its processor architectures year-after-year in its tick-tock cycle. Not AMD.

If you want to help AMD, put up a "Donate to AMD" PayPal button here and request it to be a sticky by the admins here. I'm sure AMD wouldn't mind the extra monetary help if you want a better faster and performing processor.

Give AMD time, and maybe just maybe all of their investments will finally payoff.

 

[Cool Vibrations]

AMD: why allow a monopoly on performance by your direct competitor in all categories while fighting for market share in a market that takes the performance crown of each generation extremely seriously?

Pretty hard to compete with a larger company with 13x more revenue that can concentrate most of their efforts on R&D...

 

[/dev/null]

I think OP defines enthusiast too narrowly. There are plenty of reasons an Enthusiast would choose AMD.

 

[Ticker305]

I think OP defines enthusiast too narrowly. There are plenty of reasons an Enthusiast would choose AMD.

exactly, I know plenty of car enthusiast that can afford exotics and factory race cars, but find that boring and prefer the challenge of turning a turd into a skidmark maker. Anyone can buy an Intel system and pull a monster benchmark score, but taking something like an FM2 chip and clocking it 5mhz+ to hang with an i7, gives some people a form of satisfaction that just can't be described.

I currently can't afford to take gambles, but if I had money to play with and gaming wasnt my main priority for a desktop, I'd definitely buy an AMD chipset or two to play with and put them to work on non single threaded apps if I even had a semi serious use for them.

 

[Windreaper]

+1

In my experience, enthusiasts aim for the price-performance because it's simply more fun that way.

 

[thomasward00]

exactly, I know plenty of car enthusiast that can afford exotics and factory race cars, but find that boring and prefer the challenge of turning a turd into a skidmark maker. Anyone can buy an Intel system and pull a monster benchmark score, but taking something like an FM2 chip and clocking it 5mhz+ to hang with an i7, gives some people a form of satisfaction that just can't be described.

I currently can't afford to take gambles, but if I had money to play with and gaming wasnt my main priority for a desktop, I'd definitely buy an AMD chipset or two to play with and put them to work on non single threaded apps if I even had a semi serious use for them.

Most Adults don't live on video games, they want decent performance at a good price. Just like myself I'm 39 and i'm about to build an all around workstation, I still love building my own systems and only occasionally game, the days of me spending 1500 on a gaming computer are over, in todays world there is simply no need.

 

[pelo]

You seem to think Intel cares about the enthusiast market as well...

Most PCs sold are laptops. Most money made from PCs sold are laptops. The highest margins are in servers. Most desktops sold are OEM built. Very few desktops are actually built by the enthusiast.

Simply put, AMD no longer has an enthusiast platform because there's no money in enthusiast platforms.

 

[Acer_Sheep]

Sometimes I don't understand what is that enthusiast platform everyone is speaking about. Was either AMD or Intel ever focused on enthusiast platfrom? The only "enthusiast" cpu from AMD was Athlon 64 FX, from intel there were also only the few in history mainly P4 extreme, C2 extreme and 1366/2011-based CPUs. Which those were greatly overpriced and offered poor performance in relationship to the rest of their fleet and lack sales alot. The large companies like these, are not focusing on some minority of hardcore overclockers and gamers. If either AMD or Intel would be "enthusiast" platform, I'm sure they are bankrupt for good time now.

 

[buttons]

We may each have our own ideas what enthusiast is. To me its simple, its taking a product and tweaking, tuning, etc to extract is much performance from an item as possible. Maybe im more of a "blue collar" enthusiast but the intel extreme chips have always been a joke to me. You get pretty much the same CPU thats binned slightly higher but costs 5x as much? the whole idea offends me.

To me it's performance is tied in with its price and features. You dont take a Dodge Ram and pit it against a corvette. They have different price points, different style, different uses etc. The corvette is clearly faster, better on gas, but when it comes to towing...

AMD's flagship wont be competitive to intels flagship, its been that way for a long time and continue to be that for a great long time ahead. However the 8350 at its price point will be a lot of fun to tweak, tune, etc . It wont beat the I5 in IPC or energy efficiency but it does extremely well in heavy multithreaded tests and is slightly cheaper.

 

[Gman1979]

As someone with a Masters in CS, I can tell you AMD is far from over. They are definitely lacking on the desktop side but have you seen the multithreaded benchmarks for even the original bulldozers? These things will be finding their way into many supercomputer configurations we will see coming up in the next couple years. The main thing hurting them right now is power usage, which is going to require getting to 28nm and new transistor arrangements to get competitive with Intel again in the power usage realm. The margins for enterprise hardware make desktop margins look anemic, so making chips aimed at that segment is not a bad move at all. If it works out it will definitely help recapitalize them and hopefully help them repair part of the brain drain they suffered a few years back.

Also keep in mind, Technically Sun's later Sparc cpus were "slow" in single threaded apps too. But throw them at a benchmark that could use threads effectively and they were wonderful at that use. The new designs from AMD have this same trait. I threw a cheap database server together recently using a 8150 and it was a definite step up from the x6 cpu before it once I tuned it to make use of the extra threads.


And one more thing, how can you argue with the Trinity platform? The price/performance ratio for these systems is very very nice both for desktop and mobile. It opens up the doors to the average joe being able to actually play games, watch bluerays, covert movies all on an integrated gpu matched to a decently powerful cpu without having to know anything about buying a discrete card. I never thought I would see the day a platform this powerful was considered entry level back when I was trying to play Quake 3 on a rage3d pro igp at less than 20 fps at 800x 600 lol.

 

[PcGuy78]

Too many people on this forum speculate and assume. Intel is not as far ahead as you think and focusing on desktops is a thing of the past. As far as not being "enthusiast"--ur trolling stop plz. I7 this i7 that all I hear is i7--it's a quad based on old architecture and better be the best @x86 computing. Personally, when I hear the debate about one being better than the other it demonstrates ignorance. Each"review" is set up with these "standard" OC settings on systems which seriously fail to extract the most from AMD based chips. Tired of seeing weak-ass 200/2k/2k setups on these chips like 20x200 is an oc. That's just adjustment of the multi. WIder buses, IMC tweaks(driving settings etc), and proper mem freq all yield more performance. I know this to be true because my 1st run of fritz chess(ever) and my 3.91Ghz oc is 200 points higher than the posted score of their 3.91Ghz rec for x4 965's. AMD is more enthusiastic than you're crediting. I'm not here for a rant or back and forth. Everything said is fact in this reply. INTEL has one product the "i" cores. And how many "i" core sytems run "AMD" gfx--laughable.

--

 

[/dev/null]

I don't think you even have to overclock to be an enhtusiast.

I consider myself an "enthusiast" and run a Pentium E5800 desktop with 8G ram and a 1090T with 32G ram.

I love messing with hardware & software & making them do new things (such as the vm server, pbx, mailserver, etc that I run). I am hoping to find some time to work on lan/multicast tv.

 

[Windreaper]

And if you think about enthusiast platform in terms of tweaking, AMD has full virtualization and unbuffered ECC RAM support down to the cheapest desktop processor.

 

[Dan_D]

In simple Terms, AMD has never had the financial resources to compete with Intel and even when they had the products to do so they were burdened by mismanagement and generally poor decision making which led ultimately to the state they are in now.

Time for a history lesson. (Again)

AMD was once a builder of Intel CPUs. Basically Intel was a fabless design firm. At some point Intel decided to purchase / build their own fabs and took everything internaly. After that AMD began making knock-off clones of Intel's CPUs. I believe they changed just enough to stay out of trouble. Usually it was the cache design which changed significantly vs. Intel's CPUs. The high point of this came to a head in the 386 and 486 days when Intel took AMD to court over the name "486" as AMD called their CPUs "486's" as well. The court ruled that numbers could not be copywritten. At that point Intel made the decision to change their marketing strategy and the next CPU from them wouldn't be a 586 but rather "Pentium."

AMD's 486 line actually stretched farther than Intel's as they had become fairly proficient at making them. While not clock for clock a better choice their clock speeds did outstrip Intel's. Unfortunately by the time they did this the Pentium processors were gaining market penetration and clock for clock their IPC performance was probably twice that of a 486. Despite having the FDIV bug which necessitated a recall, Intel handled the recall well and began their aggressive "Intel Inside" marketing campaign in an effort to restore their damaged reputation.

After the introduction of socket 5 motherboards (many were built with P54/P55 compatible sockets which were essentially 5/7) AMD would introduce their K5 as a direct competitor to the Pentium P54C CPUs which came in 75, 90 and 100MHz varieties. Unfortunately AMD couldn't match Intel's frequencies at that time. This is when AMD introduced (along with Cyrix) their "PR" system or "performance rating" system which was actually based on Intel's clock speeds. A K5 PR75 was equivalent to a Pentium 75 even though it's clock speed would be lower. By this time Intel had already lept far ahead of everyone else's manufacturing capabilities and no one else in the CPU industry could match Intel's clock speed ranges. Though clock for clock AMD and Cyrix offerings were sometimes superior, they didn't match the clocks of their Intel counterparts and their designs were weak in other areas. Namely floating point performance.

A little trivia: AMD's "K" was short for "Kryptonite" which is the only thing that could hurt Superman. This is how they viewed Intel at the time and hoped the K series CPUs would be like Kryptonite to Intel. Sadly the K5 fell well short of their expectations.

At that time the NextGen system's Pentium alternative the Nx586 was actually competing well in most tests (until the introduction of the Triton chipset by Intel) but naturally had the weakness of not having a built in MathCo processor initially. It was optional and sold as the Nx587 and could be added to the board which was proprietary anyway unlike AMD's offerings and used their own NxVL chipset. Later variants of the Nx586 integrated both the Nx586 and Nx587 in one package.

AMD purchased NextGen and gave them a lot of resources to work with. NextGen was already well into the design of the Nx686 and had begun working on the Nx786 as well. The Nx686 would be finished as the K6 processor and while not necessarily a Pentium killer it was a good design. (Unfortunately many Super7 boards used by the K6 II and K6 III were nothing short of trash.) The FIC VA503+ still makes the top of my list as the worst motherboard of all time. Even the NVIDIA reference design 680i SLI chipset based boards aren't as bad. I think the K6 and it's successors were generally good CPUs but suffered from having a shitty platform. I am not certain if Super-7 was a good move in all honesty. NextGen's NxVL chipset was actually pretty good. I think they should have evolved that.

For K7 AMD took advantage of the fact that Digital Equipment Corporation was winding down semi-conductor design and manufacturing and hired a bunch of engineers who worked on the DEC Alpha series CPUs. It was this direction which allowed K7 and conversely K8 to succeed to the degrees they did. Unfortunately AMD didn't get to keep all their engineers. One former AMD engineer made mention of management changes and changes to project direction which hurt them in the long run as it drove most of the talent away from the company. Buying ATi has never (as of yet) proven to be the smart move on AMD's part. At least not yet though we are seeing some benefits of this in regard to the Llano and Trinity APU's and their decent integrated graphics.

AMD spun off it's manufacturing facilities which I thought was a good move. It has it's downsides but AMD simply doesn't have the money to keep up with process technologies the same way Intel does.

The reason AMD isn't an enthusiast processor manufacturer today comes down to a history of bad decision making. Truth be told I'm not even sure that the Athlon 64 and X2 were all that good. The integration of the memory controller was obviously huge and something Intel avoided for a lot of reasons, but I think the main problem was is that Netburst was just that bad. So the Athlon was really well timed more than anything. After that I think AMD underestimated Intel's ability to build CPUs. Core 2 Duo knocked the crap out of the Athlon X2 and it didn't even have an onboard memory controller. I don't think AMD ever managed the company the way they needed to, especially in terms of R&D to keep their lead established with the Athlon 64. Basically they fell several rungs down the ladder in that one misstep with Core 2 Duo and have never been able to catch up.

You don't design a mediocre architecture and then hope to improve it over time. And I don't think AMD set out to do that but AMD is either looking too far ahead of themselves or they simply miscalculated with the direction Bulldozer should take. Essentially they bet on software becoming massively parrallel and sought to achieve a processor which was good for that. Possibly due to having the server market in mind I don't know. They may have decided that they could do only one and concentrate on the server market but the sad fact is that AMD's processors as far as the desktop goes have practically been a joke for the last 4 major product cycles. Phenom, Phenom II, Zambezi and now Vishera.

Either AMD is fucking brilliant or they are retarded fools who will run the company into the ground. Right now it looks like the latter. I think AMD sees the desktop and content creation workstations, and gaming rigs being a total niche and is trying hard to become the staple of the server market and mobile markets. Right now that's where the money is. And I have to admit, for certain server segments Bulldozer makes sense. Trinity and Llano make sense in certain applications.

Intel has the resources to tackle all these markets at once. AMD doesn't. They may simply be choosing their battles wisely and positioning themselves for the future. The unfortunate reality is that Intel can afford mistakes and AMD can't. If AMD's wrong about the direction the market will go in they are through.

And if you think about enthusiast platform in terms of tweaking, AMD has full virtualization and unbuffered ECC RAM support down to the cheapest desktop processor.

Two features most gamers and enthusiasts hardly give a shit about. They want gaming performance and for the most part if you have that, you have everything else the home user could want in spades. Machines for virtualization in the home are probably even more of a niche than gaming PC's. ECC RAM is overrated and most people won't opt for it even if the platform supports it.

 

[PcGuy78]

There's nothin mediocre about AMD, period. In computing INTEL has ALWAYS trailed AMD in innovation whereas INTEL leads in bottom line performance it lacks in originality and forsight. Each revision of AMD chips improved upon its predecessor and led in innoveation beginning w/the K6 which in fact destroyed pIII in every mark and initial PIV offerings. It wasn't until INTEL introduced HT that they even gauged nxt to AMD chips. It's easy to replicate tech but how many innovations outside of SSe4 and HT are INTEL responsible for. In fact, the AMD forums, across the globe, consistently have interesting and new findings about AMD architecture. INTEL boards are so bland. there is no "excitement" there because those users "expect" performance for little effort. This whole AMD is inferior is an opinionated stance and if you're judging by benches I suggest grabbing a piece of any questionable tech and actually use it, oc and perform your own benches. It's ez to see that INTEL is only playing it safe and waiting to see how the public reacts to "the unknown". Wasn't soon after the success of AMD's APU's that INTEL shifted focus and began seriously developing their "4000" line. There's no clear winner here a fact further supported by systems containing INTEL CPU's and AMD GPU's.... AMD competes well against the TWO major vendors in PC's by itself. The fact is that most people are not going to overpay for 5-7fps or 3secs diff in encoding unless they're already used to wasting money. Demonstrate both an AMD and INTEL setup nxt to each other running an identical 3D app or movie without telling people which is which. There's no distinguishable differences and that's the major point never discussed. All hype over benches when "real-world" is what "really" matters.

 

[Tsumi]

There's nothin mediocre about AMD, period. In computing INTEL has ALWAYS trailed AMD in innovation whereas INTEL leads in bottom line performance it lacks in originality and forsight. Each revision of AMD chips beginning w/the K6 which in fact destroyed pIII in every mark and initial PIV offerings. It wasn't until INTEL introduced HT that they even gauged nxt to AMD chips. It's easy to replicate tech but how many innovations outside of SSe4 and HT are INTEL responsible for. In fact, the AMD forums, across the globe, consistently have interesting and new findings about AMD architecture. INTEL boards are so bland. there is no "excitement" there because those users "expect" performance for little effort. This whole AMD is inferior is an opinionated stance and if you're judging by benches I suddest grabbing a piece of any questionable tech and actually use it, oc and perform your own benches. It's ez to see that INTEL is only playing it safe and waiting to see how the public reacts to "the unknown". Wasn't soon after the success of AMD's APU's that INTEL shifted focus and began seriously developing their "4000" line. There's no clear winner here a fact further supported by systems containing INTEL CPU's and AMD GPU's.... AMD competes well against the TWO major vendors in PC's by itself. The fact is that most people are not going to overpay for 5-7fps or 3secs diff in encoding unless they're already used to wasting money. Demonstrate both an AMD and INTEL setup nxt to each other running an identical 3D app or movie without telling people which is which. There's no distinguishable differences and that's the major point never discussed. All hype over benches when "real-world" is what "really" matters.

Intel: first with micro-OP cache, which AMD will be implementing something similar in Steamroller
Intel: first with 256-bit wide FPUs
Intel: first with integrated northbridge on CPU
And I'm pretty sure there are more..

As much as I love AMD and want them to be successful, let's be realistic. Both companies have their fair share of unique innovations.

 

[Dan_D]

You are so wrong it's sad.

There's nothin mediocre about AMD, period. In computing INTEL has ALWAYS trailed AMD in innovation whereas INTEL leads in bottom line performance it lacks in originality and forsight.

However you want to define innovation is irrelevant. Intel has had better performing processors at every turn for at least 13 of the last 20 years. As for foresight? One is bleeding cash and the other is paying out stock dividends. Bullshit spewed from the anus's or mouths (I get confused when talking to marketing people) of AMD's PR team about "native quad core design" and processor elegance is meaningless when it doesn't perform. And often times those chips that were more "elegenat" used more power, were manufactured on a larger process and thus wasted more materials. Statements about elegance are meaningless when the facts don't back that up.

Each revision of AMD chips improved upon its predecessor and led in innoveation beginning w/the K6 which in fact destroyed pIII in every mark and initial PIV offerings.

Each Intel CPU improved on it's predecessor. Pentium P54C improved on P5's 60/66MHz processors. P55C added MMX and lower power use. Pentium Pro improved on Pentium's design and was so far ahead of it's time that it would be revisited for generations to come as the basis for their designs. How is that not innovation?

And you've got your history wrong. K6 destroyed Pentium MMX procssors and even the Pentium Pro to a large degree. (Mostly due to clock speed advantages over the aged Pentium Pro.) Pentium II, not so much. And again the Super-7 platform was trash. Chipsets with AGP compatibility problems and driver issues out the ass do not make for a good overall system. K6 II and II competed with later PII's and PIII's but it didn't destroy them as some AMD fanboys want to believe they did. They won some and lost others.

It wasn't until the introduction of the Athlon that the Pentium III really started to falter. Early Pentium IV's were worse per clock than Pentium III's and didn't fair well against the Athlon. Even non-HT equipped Pentium IV's based on the Northwood core often competed well but still weren't as good as some Athlon CPUs. The Athlon 64 dominated and Intel couldn't remotely match it at anything but video encoding and even then not by leaps and bounds.

And even then the original Athlon suffered from motherboards with bad voltage regulation which led to FIRE. I had burnt up boards from 6 different manufacturers on my bench at any given time. They were shit. At least the ones used in OEM machines were.

It wasn't until INTEL introduced HT that they even gauged nxt to AMD chips. It's easy to replicate tech but how many innovations outside of SSe4 and HT are INTEL responsible for.

No, it is Hyperthreading that greatly improved Pentium IV's lead over earlier Athlon XP chips. It also left Intel systems often feeling more responsive while multitasking compared to AMD's Athlon XP and 64 CPUs. And while AMD did give us X86-64 and integrated memory controllers, Intel has given us so much more. USB, PCI, PCI-Express, various flash memory technologies, manufacturing processes, wireless power, Hyperthreading, integrated PCIe controllers and I/O hubs, Thunderbolt and many more standards and technologies than I can't easily recount.

Try again.

In fact, the AMD forums, across the globe, consistently have interesting and new findings about AMD architecture.

Yeah reading the processor erratum for AMD chips is scary and fun at the same time.

INTEL boards are so bland. there is no "excitement" there because those users "expect" performance for little effort. This whole AMD is inferior is an opinionated stance and if you're judging by benches I suddest grabbing a piece of any questionable tech and actually use it, oc and perform your own benches. It's ez to see that INTEL is only playing it safe and waiting to see how the public reacts to "the unknown".

I'm sorry but I review boards for a living and I see a lot more exciting boards on the Intel side. Boards like the Gigabyte Z77X-UP7 and ASUS Rampage IV Extreme are far more interesting than anything I've seen on the AMD side. AMD's processors lose in virtually every performance metric, real or synthetic. They also don't clock as high and are slower per clock and require more effort to tune. That's not being awesome, that's not being the best, it's being #2. And in a market with only two competitors that also means they are the bottom of the barrel as well.

I don't dislike AMD but your post is either trolling or you are completely ignorant of the facts. AMD is not doing all that well right now and aside from certain markets, their product line doesn't look good. Mainly their desktop CPUs which are frankly pretty pathetic right now. This is nothing new as this has been the norm for most of the last two decades.

Wasn't soon after the success of AMD's APU's that INTEL shifted focus and began seriously developing their "4000" line. There's no clear winner here a fact further supported by systems containing INTEL CPU's and AMD GPU's.... AMD competes well against the TWO major vendors in PC's by itself. The fact is that most people are not going to overpay for 5-7fps or 3secs diff in encoding unless they're already used to wasting money. Demonstrate both an AMD and INTEL setup nxt to each other running an identical 3D app or movie without telling people which is which. There's no distinguishable differences and that's the major point never discussed. All hype over benches when "real-world" is what "really" matters.

Actually Intel had integrated graphics in it's Lynnfield processors before AMD launched the first Llano APUs. And technically Intel has more graphics market share than AMD and NVIDIA combined. Intel sells more CPUs than AMD does by a wide margin. They in a sense sell more GPUs than AMD does. Though AMD has a much more broad product range thanks to purchasing ATI. In the CPU market AMD doesn't compete well. The APU is a great thing but as a CPU is basically sucks. I won't deny that the APU has vastly superior graphis performance compared to Intel's HD 3000 series graphics. I've used them both and there is no comparison but the innovation of doing it first isn't AMD's.

As for the part about a Pepsi challenge, we do talk about that. And in fact most of the time your correct. It wouldn't matter a whole lot for basic office use, surfing the web or watching movies. Get into gaming and heavy content creation and the picture changes and it doesn't look good for AMD.

 

[Acer_Sheep]

There's nothin mediocre about AMD, period. In computing INTEL has ALWAYS trailed AMD in innovation whereas INTEL leads in bottom line performance it lacks in originality and forsight. Each revision of AMD chips improved upon its predecessor and led in innoveation beginning w/the K6 which in fact destroyed pIII in every mark and initial PIV offerings. It wasn't until INTEL introduced HT that they even gauged nxt to AMD chips. It's easy to replicate tech but how many innovations outside of SSe4 and HT are INTEL responsible for. In fact, the AMD forums, across the globe, consistently have interesting and new findings about AMD architecture. INTEL boards are so bland. there is no "excitement" there because those users "expect" performance for little effort. This whole AMD is inferior is an opinionated stance and if you're judging by benches I suggest grabbing a piece of any questionable tech and actually use it, oc and perform your own benches. It's ez to see that INTEL is only playing it safe and waiting to see how the public reacts to "the unknown". Wasn't soon after the success of AMD's APU's that INTEL shifted focus and began seriously developing their "4000" line. There's no clear winner here a fact further supported by systems containing INTEL CPU's and AMD GPU's.... AMD competes well against the TWO major vendors in PC's by itself. The fact is that most people are not going to overpay for 5-7fps or 3secs diff in encoding unless they're already used to wasting money. Demonstrate both an AMD and INTEL setup nxt to each other running an identical 3D app or movie without telling people which is which. There's no distinguishable differences and that's the major point never discussed. All hype over benches when "real-world" is what "really" matters.

What is not in favor of AMD either is way too high media coverage and overrated intel benchmarks. If there wouldn't be all this shit in magazines and websites reviewing hardware around the world. I'm sure that 90% of those having AMD didn't even notice they are running "slower" AMD CPU. Or intel owners running faster ones, the performance differences and energy consumption differences are so small in large scale that for most users either the gamers, basic or specialized ones there would be no difference in running the processor of either one or other company. Sadly impressing peeps over highest benchmark scores has been here since day 1 and is basically important, though useless.

 

[Dan_D]

What is not in favor of AMD either is way too high media coverage and overrated intel benchmarks. If there wouldn't be all this shit in magazines and websites reviewing hardware around the world. I'm sure that 90% of those having AMD didn't even notice they are running "slower" AMD CPU. Or intel owners running faster ones, the performance differences and energy consumption differences are so small in large scale that for most users either the gamers, basic or specialized ones there would be no difference in running the processor of either one or other company. Sadly impressing peeps over highest benchmark scores has been here since day 1 and is basically important, though useless.

There are situations where performance matters. I know many AMD fans can't seem to understand that. If I'm rendering graphics for computer animation then time is literally money and doing a frame 15 seconds faster could equal hours of saved time and thus money. This thinking applies to server markets as well though AMD is more competitive there. And when it comes to gaming, there are certain situations where the difference can be pretty large. Large enough to justify the increased cost of a better CPU.

For the bulk of the masses, it doesn't matter much. But for enthusiasts and professionals it can.

 

[pelo]

There are situations where performance matters. I know many AMD fans can't seem to understand that. If I'm rendering graphics for computer animation then time is literally money and doing a frame 15 seconds faster could equal hours of saved time and thus money. This thinking applies to server markets as well though AMD is more competitive there. And when it comes to gaming, there are certain situations where the difference can be pretty large. Large enough to justify the increased cost of a better CPU.

For the bulk of the masses, it doesn't matter much. But for enthusiasts and professionals it can.

This might be a bit backwards...

AMD has a very small server market share, and most of what it has been implemented in select HPC's. They're hovering just around 5%. They offer moar coars, but it's at the cost of power consumption. That's a pretty big deal in server. The licensing issues regarding Windows server and coars also doesn't help them.

In workstation applications these 8350's do very well. They may not perform as well as a 3930K, but they also cost only ~$200. Generally speaking, workstation applications tend to be pretty well parallelized, thus the single core deficiencies don't come into play.

The 8350s are like cheaper 2011 chips. They may not offer performance that's as good, but it's still pretty good.

sub-$200, the 2 and 3 module parts have made any recommendations for i3's basically void. There's really no way I'd recommend an Intel i3 over any AMD CPU anymore. The 2 and 3 module Vishera chips will outshine them even in gaming.

For the enthusiast/gamer these chips make no sense, though. Just buy a 2500K and be done with it.

Although I hate tom's, they've taken a pretty decent approach to benchmarking recently (I believe they did a complete reworking of their benchmarking procedures last year). Instead of just posting synthetic benchmarks, they'll do synthetic benchmarks to expose the underlying architecture but focus a majority of their benchmarks on actual applications. Obviously, this is a much better approach than just posting synthetic benchmarks by themselves, which is utterly meaningless unless you're going to explain why and what the synthetic benchmark results mean.

There's more, but generally speaking the 8350 does quite well in productivity related tasks.

So, let’s try to distill all of this down into a recommendation. Recognizing that the power user community gives AMD more latitude than Intel, I anticipate a greater number of enthusiasts getting excited about FX-8350 than any of the Bulldozer-based CPUs, and rightly so. More speed, significantly improved efficiency, and a sensible price tag are exactly what I was hoping to see, and AMD delivers them all. Are you asked to make compromises? Yeah. Single-threaded performance still isn’t impressive, and power consumption remains a sore subject. But for less than $200, I can certainly see FX-8350 at the heart of a budget-oriented workstation.

Which isn't surprising, frankly. Productivity related software tends to scale very well and also be updated quite frequently, meaning those shiny new ISAs and moar coars get to stretch their legs. Power consumption, too, isn't as big a concern as it is in server either.

 

[Soarin]

Okay after reading most of this I am sad. I personally like both companies. I am rocking a phenom II x2 right now. I just can't see spending 350 on a cpu. That seems awfully high. Then again being raised poor may be talking too,.

 

[Dan_D]

This might be a bit backwards...

AMD has a very small server market share, and most of what it has been implemented in select HPC's. They're hovering just around 5%. They offer moar coars, but it's at the cost of power consumption. That's a pretty big deal in server. The licensing issues regarding Windows server and coars also doesn't help them.

In workstation applications these 8350's do very well. They may not perform as well as a 3930K, but they also cost only ~$200. Generally speaking, workstation applications tend to be pretty well parallelized, thus the single core deficiencies don't come into play.

The 8350s are like cheaper 2011 chips. They may not offer performance that's as good, but it's still pretty good.

sub-$200, the 2 and 3 module parts have made any recommendations for i3's basically void. There's really no way I'd recommend an Intel i3 over any AMD CPU anymore. The 2 and 3 module Vishera chips will outshine them even in gaming.

For the enthusiast/gamer these chips make no sense, though. Just buy a 2500K and be done with it.

Although I hate tom's, they've taken a pretty decent approach to benchmarking recently (I believe they did a complete reworking of their benchmarking procedures last year). Instead of just posting synthetic benchmarks, they'll do synthetic benchmarks to expose the underlying architecture but focus a majority of their benchmarks on actual applications. Obviously, this is a much better approach than just posting synthetic benchmarks by themselves, which is utterly meaningless unless you're going to explain why and what the synthetic benchmark results mean.

Snip....................


There's more, but generally speaking the 8350 does quite well in productivity related tasks.



Which isn't surprising, frankly. Productivity related software tends to scale very well and also be updated quite frequently, meaning those shiny new ISAs and moar coars get to stretch their legs. Power consumption, too, isn't as big a concern as it is in server either.

I like how there isn't a single Sandy Bridge-E benchmark in the mix. Vishera is doing better than I thought it would but it's still getting beaten by a processor with half as many cores just through Hyperthreading. That speaks volumes about the architecture of Vishera right there. I do understand there is some bang for your buck arguments that can be made for Vishera but my point is that there are times where this is more to computing than just getting the cheapest thing. Sometimes performance matters. That was my point. The applications you mentioned should scale even better on Sandy Bridge-E and again several of Intel's offerings still perform better than Vishera does in almost all cases.

Vishera is a nice improvement over Zambezi. I'm not denying that. Again I'm not denying bang for your buck either. But I find it interesting that bang for your buck is often the only argument being made as if performance never matters on it's own.

This isn't [F]rugal|OCP last time I checked.

 

[CleanSlate]

This isn't [F]rugal|OCP last time I checked.

lol

 

[buttons]

Okay after reading most of this I am sad. I personally like both companies. I am rocking a phenom II x2 right now. I just can't see spending 350 on a cpu. That seems awfully high. Then again being raised poor may be talking too,.

$350 ? its $199 and works on any am3+ board which you might already be using.

 

[buttons]

Vishera is a nice improvement over Zambezi. I'm not denying that. Again I'm not denying bang for your buck either. But I find it interesting that bang for your buck is often the only argument being made as if performance never matters on it's own.

This isn't [F]rugal|OCP last time I checked.

Yeah, i like to support amd but lets be real for a moment. This is an enthusiast forum which means we not going to leave our CPU at stock clocks. I fear an i5-3570k overclocked to 8350 speeds may turn the tables even in multithreaded tests.

 

[Dan_D]

Yeah, i like to support amd but lets be real for a moment. This is an enthusiast forum which means we not going to leave our CPU at stock clocks. I fear an i5-3570k overclocked to 8350 speeds may turn the tables even in multithreaded tests.

Overclocked the 3570K and 3770K are going to be faster than Vishera at most desktop applications. Content creation applications which are designed to be multithreaded and scale past 4 cores show Vishera in a better light than your average desktop applications do. Even in a best case scenario Intel's 4c/8t CPUs still overtake an 8 core Vishera processor in these applications. The benchmarks above show the 3770K beating Vishera in almost all those multithreaded tests with a clock speed deficit. (A clock speed deficit of at least 300-400Mhz accounting for both CPUs turbo frequencies.) And again the Intel's will overclock at least as well as Vishera does and probably better. And note that we aren't even getting into Intel's 6c/12t and 8c/16t offerings.

Let's be honest here. Intel's should overclock as well or better than Vishera, have higher IPC, lower power consumption, less heat, more memory bandwidth and in most respects a better and more advanced motherboard platform. Arguments against these issues which downplay the importance of performance, power consumption, etc. are all signs that some people don't want to face the truth. Again I do realize that when cost comes into play the picture does change and AMD processors can be compelling at specific price points in certain circumstances. I am aware of that but I'm pointing out that bottom dollar PC builds aren't the only consideration. And frankly I'm not sure when the [H] went so oft. It seems lately that people like me are often talked down to like we are sub-human for spending a little money on a hobby and told that we are wasteful and that performance doesn't matter or something. I'm an enthusiast first and foremost and AMD doesn't really look compelling for anything but budget builds. I realize that that's where some people are at in their lives and that's what they have to work with. I've been there. So I am aware of where some of you are coming from.

Again I'm not saying that these products never make sense or shouldn't be used, or even that they are all bad. But come on, a quad core CPU beating out one with 8 real cores with a clock speed deficit is pretty fucking sad. AMD's best wouldn't be competitive at all right now if it weren't priced so low.

 

[CleanSlate]

So I'm not saying that these products never make sense or shouldn't be used, or even that they are all bad....


But come on, a quad core CPU beating out one with 8 real cores with a clock speed deficit is pretty fucking sad. AMD's best wouldn't be competitive at all right now if it weren't priced so low.

Bottom line is that people want to see a company with the best performance available or at least a competitive edge in some areas worth their while.

Where's the edge?

 

[Dan_D]

Bottom line is that people want to see a company with the best performance available or at least a competitive edge in some areas worth their while.

Where's the edge?

I know what they want. They want a return to the glory days when the Athlon 64 reigned supreme and put Intel on the the defensive. Not only that but they want AMD to be more than they are in the CPU market. They want that one period of Athlon dominance to be something other than a fluke or the product of chance. This is because people personify companies for some reason and become emotionally invested in brands. I always buy whatever I feel is the best product. I temper that within a budget if there is a budget attached to whatever it is I'm doing. I'll buy a $1,000 CPU if it was something to offer that I think is worth while. When it doesn't I won't. And actually the Core i7 980X was the only one I felt was worth the cost. It offered two more cores and great overclocking.

I might want a Mercedes but I'm working with a Ford or Chevy budget on that. So I don't buy Mercedes or BMW's. I suspect some people feel the same about their CPU purchases and I do get that. But I don't hate higher end cars or the people that buy them, nor do I think they are bad products just because I can't afford them. I also do not pretend that my GM or Ford products are in the same league as a BMW or Mercedes is.

 

[Tsumi]

Bottom line is that people want to see a company with the best performance available or at least a competitive edge in some areas worth their while.

Where's the edge?

So because AMD does not have the absolute fastest desktop processor, they're going to fail. That's what you've been trying to say all along, right?

You are grabbing onto anything that furthers your point, whether or not it is in or out of context. You are also completely ignoring any counterarguments.

AMD will not fail if they don't have the absolute fastest desktop processor, or at least one that competes with it. What will cause them to fail is the failure to turn a profit. There's only one way of this happening: costs of researching and producing their chips is greater than what the chips earn. AMD priced their processors low in hopes of producing greater volume, which would offset the lower profits per processor. If this does not happen, then they will fail. Not because they don't have the fastest desktop processor, which really is only a small part of the market.

 

[Soarin]

$350 ? its $199 and works on any am3+ board which you might already be using.

Sorry I did not complete my post. Intel wise I cant see spending 300-350 on a CPU

 

[Dan_D]

Sorry I did not complete my post. Intel wise I cant see spending 300-350 on a CPU

You don't have to. The 3570K can be had for $229.99 on Newegg or around $199.99 at Microcenter if you have one near you. The applications where Vishera might be a better option are few and far between for the home user or gamer. The 3770K is $329.99 on Newegg or can be had for around $299.99 at Microcenter. Though it's $300 it won't be beaten by Vishera at hardly anything.

 

[Acer_Sheep]

There are situations where performance matters. I know many AMD fans can't seem to understand that. If I'm rendering graphics for computer animation then time is literally money and doing a frame 15 seconds faster could equal hours of saved time and thus money. This thinking applies to server markets as well though AMD is more competitive there. And when it comes to gaming, there are certain situations where the difference can be pretty large. Large enough to justify the increased cost of a better CPU.

For the bulk of the masses, it doesn't matter much. But for enthusiasts and professionals it can.

I didn't speak about enthusiasts at all, but I meant there are alot of professionals who don't see into technical stuff about computers and they bought intel as well because intel is more recommended and their CPUs win benchmarks, intel is also more profitable because they have more fabs and OEMs are more likely to go for intel CPUs because intel can fulfill larger orders while AMD can't. So what AMD did forget is that feedback is also important, otherwise they would not pull the second gen-FXes and destroy their reputation completely throughout 2011 and 2012, yes their failure.

Let's be honest here. Intel's should overclock as well or better than Vishera, have higher IPC, lower power consumption, less heat, more memory bandwidth and in most respects a better and more advanced motherboard platform. Arguments against these issues which downplay the importance of performance, power consumption, etc. are all signs that some people don't want to face the truth. Again I do realize that when cost comes into play the picture does change and AMD processors can be compelling at specific price points in certain circumstances. I am aware of that but I'm pointing out that bottom dollar PC builds aren't the only consideration. And frankly I'm not sure when the [H] went so oft. It seems lately that people like me are often talked down to like we are sub-human for spending a little money on a hobby and told that we are wasteful and that performance doesn't matter or something. I'm an enthusiast first and foremost and AMD doesn't really look compelling for anything but budget builds. I realize that that's where some people are at in their lives and that's what they have to work with. I've been there. So I am aware of where some of you are coming from.

You must understand that people don't want to face the truth, but you don't want to face it either, for whatever reason, wastefully investing money into top end hardware for the sake you just want it or can afford it, is illogical same way as is being attached to particular brand or company. And people were always doing this, doesn't matter if it is in computers, it is everywhere you get rock solid fans and supporters of companies in transportation, industrial stuff, computing, musical instruments and art and the list goes much longer.
In other hand the situation on the forums, and not just at [H] is very much reflection of current situation on the market and in the global economy at all. First of all budget builds in the past were able only to work with office, internet and other very basic needs. They were not able to run almost any games and particularly also no media. Today budget builds are capable of vastly more and even the cheapest computers on the market can be freely used for multitasking and amateur working with multimedia and as well as quality casual gaming and even discrete graphics is not needed anymore because in last 2 years the IGP performance has rose by 60% if not even more when considering APUs.
With minor upgrades, budget builds are able to be hardcore overclocked and game almost as the enthusiast ones.Which infact narrowed the entire enthusiast segment to tiny piece of nothing and in that case, unless person does something really hardcore such as animating or doing scientific calculations etc, I don't really see any reasons or pros of buying something like sandy bridge E for my home computer while I can have the same way stuff done for $700 - 800 cheaper CPU and have same fun with gaming, overclocking it or tweaking it in any other way.
The other reason is, the longer we go, the economic situation in north america and europe is worse and also life is harder in that case, so the investments towards optional things such as enthusiast hardware are downed by this as well.

 

[Woe]

AMD went to shit once Bulldozer turned out to be a flop. Llano/Trinity is where AMD shines now.

 

[pelo]

I like how there isn't a single Sandy Bridge-E benchmark in the mix. Vishera is doing better than I thought it would but it's still getting beaten by a processor with half as many cores just through Hyperthreading. That speaks volumes about the architecture of Vishera right there. I do understand there is some bang for your buck arguments that can be made for Vishera but my point is that there are times where this is more to computing than just getting the cheapest thing. Sometimes performance matters. That was my point. The applications you mentioned should scale even better on Sandy Bridge-E and again several of Intel's offerings still perform better than Vishera does in almost all cases.

Vishera is a nice improvement over Zambezi. I'm not denying that. Again I'm not denying bang for your buck either. But I find it interesting that bang for your buck is often the only argument being made as if performance never matters on it's own.

This isn't [F]rugal|OCP last time I checked.

But it isn't gnoreeverythingelsethatdoesntrevolvearoundgaming|OCP.

I mentioned the SB-E in the post, but it's also much more expensive. If you think about it, what normal usage would the chip be insufficient in other than slip behind in gaming?

If you're going to benchmark then benchmark with actual applications that people use. That's exactly what Tom's did. Using synthetic benchmarks alone is akin to measuring penis sizes according to how quickly it runs cinebench but avoiding a ruler.

But I find it interesting that bang for your buck is often the only argument being made as if performance never matters on it's own

Because it's about both price and performance. I'm not buying a 3960X because it's the best chip on the planet because it costs a bagillion dollars. Likewise, there's clearly an advantage in opting to go with a 8350 over a 3770K if you can spend the ~$120 on a GPU or RAM or what have you. Ultimately you're going to get a better rig if your cash is evenly spread out amongst the parts.

I still think the 8350 is the worst pick of the bunch, imo. It has the smallest niche to fill and competes with the unlocked quads. That's a tough sell. f I were building a rig today I'd avoid the 4-module parts and go with a 2500K. But can you say the same for the 2 and 3 module chips? Those are clearly better positioned.

I don't think AMD can survive 2013, but that doesn't mean I'm willing to ignore actual worthwhile benchmarks and reach the conclusion that the Vishera chips aren't a good buy.