2.35ghz? Do they clarify if that's when running PS5 Pro optimized code (60CUs) without enabling the High CPU Frequency Mode or is it actually the clock speed of the Ultra Boost Mode (36CUs) for unpatched games?
Ah, alright.
Max boost clock? Was the clock in the PS5’s GPU referred to as such?
Kinda. The 2.23ghz figure was always refered to as the peak of a variable frequency, just like the CPU's 3.5ghz, due to the utilization of AMD SmartShift.
If we're lucky, it'll be like the regular PS5 and actually hit those frequencies 99% of the time. If it's the case, we're looking at a slightly more powerful console than expected. That would be very nice.
Is it 3,840 x 4 x 2.35?
Like
HeisenbergFX4
has been repeating for several months now. 3080/4070/7800 XT tier of performance in rasterization.
I have a 4070, i'm curious to see what they can get out of the perfomance on the PS5 Pro.Like
No. they doubled the cache... just not on the level that some of us predicted. That is, we expected the L2 cache to be doubled. They doubled the L1 and L0 cache instead. That should tell you where they feel/know their bottlenecks are.
4070 raster performanceLike
I have always thrown out there real world performance not so much brute force numbersLike
Kepler is not always accurate with his info, you can go and check for yourself, the tweets are still there.
With that being said, so far these are the SE count of RDNA4.
Navi44 - 16WGP (2SE, 32CU, 64ROPs)
Navi48 - 32WGP (4SE, 64CU, 128ROPs)
So the WGP per SE is still the same as RDNA3.
8WGP or 16CU per SE.
PS5 having 4SE is more probable than 2SE.
2SE with that many CU seems to be inefficient as well.
With higher end RDNA4 going by Navi44/48, it appears to be still 8WGP or 16CU per SE.
And it seems each SED houses 2 Shader Engines.
Navi41
1SED = 2SE/16WGP(32CU)
6SED = 12SE/96WGP(192CU)
1AID = 2×64bit or (4×32bit GDDR PHY)
2AID = 4×64bit GDDR PHY = 256bit
Navi40
1SED = 2SE/16WGP(32CU)
9SED = 18SE/144WGP(288CU)
1AID = 2×64bit or (4×32bit GDDR PHY)
3AID = 6×64bit GDDR PHY = 384bit
The concept of 1SED having 2SE is the same as Zen4C CCD having 2CCX.
Here we can see the size of 2SE.
So its actually better than PS5 clock. Nice.
2x more L0 and L1 caches. Not surprising (I was 100% expecting this) with the new CU structure (X1X is heavily L1 cache starved and has likely only 128kb L1 cache like PS5) and is likely where the better bandwith efficiency comes from. I was totally expecting some high max boost clock like this.
I wouldn't read too much into this. As of now at least, the applications of Mesh Shaders and Primitive Shaders seem to be identical. Remedy and Ubisoft Massive didn't even mention Primitive Shaders, they said Mesh Shaders, full stop, and this included the PS5.
*3090 more accurately, as I had mentioned.Like
It was always just a question of clock speeds which I'll remind folks isn't 100% confirmed until Sony releases the official specs themselves. I've said many times that the clocks weren't going to be lower than the base system. Absolutely min was going to be 60 CUs with dual issue (RDNA3+) with no clock increase (2.23 Ghz) = 34.3 TFLOPS. I was hoping we'd get at least a 10% clock bump which would put it at 2.45 Ghz (~38TFLOPS). I'm still really curious why there isn't a fixed clock increase for either the CPU or GPU assuming it's using RDNA 4 as rumored
.
Another point to setup realistic expectations. Note that the average FPS at native 4K for even the lower end of these estimated cards (4070/6800XT) is comfortably above 60. That's native with no super resolution bumps and ultra settings which console games typically do not run at. So it's absolutely within the realm of possibility that most games will be able to run at native 4K/60fps on the Pro with these reported specs.It was always just a question of clock speeds which I'll remind folks isn't 100% confirmed until Sony releases the official specs themselves. I've said many times that the clocks weren't going to be lower than the base system. Absolutely min was going to be 60 CUs with dual issue (RDNA3+) with no clock increase (2.23 Ghz) = 34.3 TFLOPS. I was hoping we'd get at least a 10% clock bump which would put it at 2.45 Ghz (~38TFLOPS). I'm still really curious why there isn't a fixed clock increase for either the CPU or GPU assuming it's using RDNA 4 as rumored
I'd just remind folks that 34-36 TFLOPs with 60CUs is ~6800XT level. That is in the 80-90% range for a real world FPS increase in games (assuming GPU bound). Nearly the 2x raster target that had been rumored for a while now:
Now if someone could explain the +45% "rendering performance" metric....?
Sony APIs for past 3 gens are basically just mirroring GPU native instruction-set, it's as thin and as close to the hw as you can get (the only way to go closer is the PS2 way where they just dumped instruction set to developers and make them write machine code directly - yes, GPU programming on PS2 was literal machine code - there were no higher abstractions or even assembler for it).
I wouldn't use another-platform benchmarks to infer relative positioning of two console GPUs, no matter how 'comparable' they seem on paper. There's tons of counter examples showing it makes for a poor methodology for comparison basis.
Interesting,the RX 7700 only has 2MB of L2 cache though it has like 8x infinite cache.
I've been ignoring this since I saw it. Not sure what it refers to but in terms of raw performance, I expect the Pro to perform 70-80% better than the regular PS5, which would put it in the ballpark of a 3080.
