Qualcomm Snapdragon X Elite Put Through Graphics Tests, Beats AMD Radeon 780M iGPU in 3DMark

erek

erek

[H]F Junkie
Joined
Dec 19, 2005
Messages
10,929
Speaking of Pat Geislinger of Intel …

“In the 3DMark Wildlife Extreme benchmark, designed for graphics solutions of this class, the Snapdragon X Elite scored 39.2 FPS (average), compared to 60 FPS of the Apple M2 Max, 40 FPS of the Apple M2. The Core i7-13700H "Raptor Lake" is a 45 W mobile processor with an Intel Xe-LP based iGPU that has 96 EU. This chip scored just 22.5 FPS in this test. The surprise here is the Radeon 780M, the iGPU of the AMD Ryzen 7 7840HS, based on the latest RDNA3 architecture, with 12 compute units (768 stream processors). This chip did just 28 FPS, falling behind even the M2. The other benchmark is "Control" at 1080p with its lowest graphics settings, and here the results are fundamentally different. With "Control," we see the Snapdragon X Elite post a respectable 53 FPS, which is almost as fast as the 56 FPS by the Radeon 780M powering the Ryzen 7 7740HS, but ahead of the 43 FPS put by the Apple M2, and a whopping 145 FPS by the M2 Max.”

1698693857126.png


https://www.techpowerup.com/315245/...cs-tests-beats-amd-radeon-780m-igpu-in-3dmark
 
7000 series APU's from AMD can contain either the 680m or the 780m part....the 780m chips had a $250 premium when I last checked, but really didn't deliver $250 worth of performance improvement. Just posted for awareness, the 7000 series contains one of two iGPU's depending on their tier.
 
  • Like
Reactions: travm
like this
I hope these are genuinely good when they make their way out into the market. More potent mobile solutions that don't eat x86 amounts of sparkle pixies can only be good for the market. I'm not likely to join the Borg, but I would love to see Windows not be shit on ARM.
 
HeadRusch said:
7000 series APU's from AMD can contain either the 680m or the 780m part....the 780m chips had a $250 premium when I last checked, but really didn't deliver $250 worth of performance improvement. Just posted for awareness, the 7000 series contains one of two iGPU's depending on their tier.
Click to expand...
More than that 610, 660, 680, 740, 760, 780, or a good old Vega (Radeon Graphics)
https://www.amd.com/en/processors/ryzen-processors-laptop
 
DukenukemX said:
There are problems with this benchmark. Firstly the 780M can beat a RTX 3050 if paired with fast enough ram. Most benchmarks are done with 4800Mhz ram where the 780m likes 5600Mhz. Also this chip is only for synthetic benchmarks. Try running an actually 3D program like a game or CAD program.
Click to expand...
If you give it a lot of power, those benchmark are made a 23 watt I think.

If they are those:
https://www.anandtech.com/show/2111...ormance-preview-a-first-look-at-whats-to-come
They were made on a asus rog zephyrus G14 (GA402XV)

Which seem indeed on 4800mhz memory:
https://rog.asus.com/ca-en/laptops/rog-zephyrus/rog-zephyrus-g14-2023-series/
 
pendragon1 said:
seems like a long way of saying "he's right"....
Click to expand...
DId not know before starting and do not know if the article is really the same as the anandtech values, cited, people that have followed all of it would know.

And obviously if it is not true that when using only 23watt the 780m would beat the 3050 in the average scenario, I do not agree with him at all that the benchmark has an issue.
 
LukeTbk said:
I do not agree with him at all that the benchmark has an issue.
Click to expand...
The problem is the benchmark is synthetic, meaning it doesn't really represent real world situations. Geekbench for example shows that Apple iPhone 15's are faster in the GPU, but in games tested the S23 Ultra is faster. This is why you should only use synthetic benchmarks for similar hardware, like an iPhone 14 vs iPhone 15. Otherwise, always use real world applications. The AMD 780m is still an AMD product, which means it'll suffer in Ray-Tracing performance compared to the RTX 3050. Meanwhile the Snapdragon X Elite has no games on Windows to make use of it's Ray-Tracing hardware. Comparing the Snapdragon's GPU to that of desktop GPU's is mostly pointless because realistically you can't run much of anything on an ARM chip on Windows. If this were Android then that's a different story, but the Snapdragon X Elite is mostly meant for Windows.
 
Lakados said:
More than that 610, 660, 680, 740, 760, 780, or a good old Vega (Radeon Graphics)
https://www.amd.com/en/processors/ryzen-processors-laptop
Click to expand...
...sweet jeebus........I was waffling on the 660 vs 680 because I thought I dreamed that up but clearly this is part of the AMD meme factory here.......I didn't realize the 7 series could literally pair with any of those depending on the part #. Basically the message intent was "it's not just 780's with the 7 series".
 
HeadRusch said:
...sweet jeebus........I was waffling on the 660 vs 680 because I thought I dreamed that up but clearly this is part of the AMD meme factory here.......I didn't realize the 7 series could literally pair with any of those depending on the part #. Basically the message intent was "it's not just 780's with the 7 series".
Click to expand...
Yeah, laptop and mobile sales are up, Desktop sales are down, and AMD is having to compete with Intel at price points that are incredibly competitive and the only way to do that is by offering a crapload of SKUs and having as little wasted silicon as possible which is where we get Intel's binning SKU hell lists from.
Laptop sales are still high volume no margin territory.
 
Big surprise, the Snapdragon Elite X is lying through their teeth.
https://it.slashdot.org/story/24/04...ing-on-their-snapdragon-x-elitepro-benchmarks
https://www.semiaccurate.com/2024/04/19/what-is-qualcomms-purwa-x-pro-soc/

"The rest of the 'disclosure' for Snapdragon X Pro was a list of features that all fall under the guise of exactly what you would expect. The rest was filled with deep 'details' like the GPU capabilities of 3.8TFLOPS. That's it. No specs, no capabilities, no nothing. It was truly pathetic. But wait there is more, or less really, with statements like it having AV1 encode and decode. Trivialities like frame rates and resolutions were seemingly not needed for such technical briefs. See what we mean by pathetic? Those 10 cores are arranged how again? That 42MB of cache is what level? Shall I go on about the bare minimum basics or do you get the point now? SemiAccurate was planning to ask Qualcomm about their cheating on benchmarks at the promised briefing but, well, they lied to us and cut us out of the pathetic bits they did brief on. We honestly would have liked to know why they were cheating but we kind of think they will do their usual response to bad news and pretend it never happened like last time. If they actually do explain things we will of course update this article as we always do."
 
Most shocking, we can imagine people betting on emulation (from openGL-vulkan, etc...) support getting better and better and betting reality end up matching the fake one.
 
DukenukemX said:
There are problems with this benchmark. Firstly the 780M can beat a RTX 3050 if paired with fast enough ram. Most benchmarks are done with 4800Mhz ram where the 780m likes 5600Mhz. Also this chip is only for synthetic benchmarks. Try running an actually 3D program like a game or CAD program.


View: https://youtu.be/MBvIbyJHzCA?si=SBhv2yMsYG5jTMVx
Click to expand...

3DMark stopped using an actual game engine over a decade ago. Since then, it's been entirely synthetic and theoretical. The old Radeon 2900XT used to out perform the 8800GTX in 3DMark but in actual games the former got beaten down hard. 3DMark performance doesn't even necessarily benefit from the same things games do. Meaning, things you might do to improve frame rates in actual games won't necessarily translate into 3DMarks.

It's a benchmark that I've argued has been useless for years. It's not really good at load testing GPU's either. There are better tools for that. It's not good at CPU load testing. I've seen systems that can run 3DMark without crashing and crash doing just about anything else and vice versa. Of course now it has the capability to guess how many FPS you'd get in certain super popular titles based on your scores and hardware configuration, but prior to that its not as if you could say 11,000 3DMarks translates to 120FPS in XYZ game. It wasn't ever that useful.

I have no idea why anyone puts stock in the test.
 
DukenukemX said:
There are problems with this benchmark. Firstly the 780M can beat a RTX 3050 if paired with fast enough ram. Most benchmarks are done with 4800Mhz ram where the 780m likes 5600Mhz. Also this chip is only for synthetic benchmarks. Try running an actually 3D program like a game or CAD program.


View: https://youtu.be/MBvIbyJHzCA?si=SBhv2yMsYG5jTMVx
Click to expand...

Hee hee.... sort of reminds you of pre-Proton (Steamdeck) days of not having many (Windows) games in Linux. I'm sure everyone is "right on this", getting their games ready for Snapdragon X.

(of course, it does depend on Microsoft's commitment on this... so, gaming could happen... maybe...)
 
You must log in or register to reply here.
Back
Top