that's a low res i'd have thought your stock is slightly overclocked compared to reference cards? isn't your card a EVGA SuperClock?
Most media concluded this slide was less a genuine expectation and more marketing fantasy given the evidence of history. Kepler is already 2012, and Maxwell depends on TSMC 22nm, which has barely even been talked about by TSMC yet so expect it 2-3 years away. We've just been designing our Kepler boxes here and I'm itching to see internal results as well
Its stock o/c at 772 core i think, so come on guys how much does a cpu play in a gpu bench? lol I need a good intel bundle 4.4ghz?
Edit: see below. I'm surprised at Pookey's single 580 score of 46fps... that's probably double what I get with a stock 480, which ain't right. Will add my results shortly.
I know it is only speculating, and as you said a 'Marketing Fantasy' but if it is going to be as fast as that graph, then that will be a massive jump in performance. As for Kepler, is it a fact that it will be the GTX 680 for 6 months, before they bring out the GTX 780 with Kepler's true performance or is this all just rumor's? From what I am guessing, Nvidia will bring out the GTX 680 to take on the HD7970. AMD will then bring out the HD7980 or HD8970 to compete with Nvidia, and then Nvidia will release the GTX 780 to take on that. (Single GPU speculating, excluding any dual GPU cards). If this is the case, then 2012 is going to be an interesting battle ground between the 2 company's. Anyhow, I hope Kepler comes out within the next 78 days, so I can get a free step-up from EVGA.
@ noiz, here's my score at 825/1650/1900 with vsync off, CPU at 4GHz (160x25) and memory at 1600MHz. There's no point running a graphics performance bench with vsync on because vsync interferes with the frame rate, lowering what the graphics card is actually capable of producing.
Post a screen of your Heaven result and I'll see how my 480 compares to your 580 with the same benchmark settings. EDIT: I get EXACTLY the same score whether the CPU is at 4GHz or 4.7GHz, so it seems that CPU clock and mem bandwidth don't affect Heaven at all.
TG, is there truth is this, have you tried downclocking and comparig benches? since you said my amd appears to be holding me back?
Yeah I can run at 4ghz to give you a comparison with HT off so it is 6cores Vs 6cores. What is your stock clocks so I can run that as well. Also memory and timings need to be the same, for better comparison. So; 4Ghz 6 cores no HT 1600 9 9 9 24 2T GPU clocks...EDIT 797mhz core 2025mhz memory. EDIT; How much AA our we using?...we need to have a set standard. 1920x1080 16X Anisotropy 4x AA High settings Tessellation normal Vsync off
yea mine are just stock 580 SC settings, dont know them off top head but should be on evga site. RAM is 1600mhz 8gb
His clocks are at 797mhz core 2025mhz Memory ^^^^ CPU 4.7Ghz 6/12 1x GPU @ 797/1594/2025(4050) 1920x1080 16X Anisotropy 4x AA High settings Tessellation normal Vsync off Heaven Benchmark
CPU 4Ghz 6/6 1x GPU @ 797/1594/2025(4050) 1920x1080 16X Anisotropy 4x AA High settings Tessellation normal Vsync off Heaven Benchmark So CPU speed with/without HT makes no differences in this benchmark, using single GPU anyhow. GTA4 might be a good benchmark for CPU speed and GPU clock speed testing.
I'm gonna re run it later today with my CPU as high as it will go (4.2ish with any luck). My score seems a hell of a lot lower than yours even considering the better graphics card. Oh and as for vsync, I forget to disable it for the first two runs Edit: What timings is your memory at by the way? Mine's currently sat at 7,7,7,24. I might slacken them and try it at 1600 to see if that makes a difference.
It's because Heaven is a DX11 benchmark so the 480 will really pull ahead - and just look at the 580 stock scores, totally demolishing the 480 at 825MHz! Crazy! As for memory I had it at 6-7-6-18. For the 4.7GHz run it was at 2000MHz 8-9-8-24, much higher bandwidth that way, and the score was identical. I think at a push I might be able to match the 580 stock performance - will give it a bash later.
I'm sure I could get closer to your score, just need more mV lol It is interesting to see how much difference there is between 2 generations of top end GPUs
