Crysis 1.2.1 Benchmark Last Change: 20/Nov/2014, 0227hrs. Whoami: Ian Mapleson http://www.sgidepot.co.uk/sgi.html http://www.sgidepot.co.uk/sgidepot/ Here I present collated results for a range of different systems for the Crysis V1.2.1 benchmark. See after the tables for extended system details. All tests done with Win7 64bit Pro or Ultimate, except for the P4 system which has to be 32bit. Key (all GPU clocks are given as 'Core/Shader/RAM'; a single number refers to just the core clock): Qu/FX = Quadro FX card 8800GT/G = Gigbyte 8800GT 512MB Zalman (Turboforce Edition), stock @ 700/1700/1900. oc'd to 775/1850/1900 on my i7 860. 8800GT/A = ASUS 8800GT 512MB Glaciator, stock @ 602/1512/1800, oc'd to 685/1725/1900 on my i7s, oc'd to 650/1608/1900 on the Athlon64 X2 6000+ 3.225GHz PC (ASUS M2N32 WS Pro, 4GB DDR2/800 RAM). 5850/S = Sapphire 5850 Xtreme 1GB, default 725/1000 core/RAM. 7970/S = Sapphire Vapor-X 7970 3GB, stock @ 1050/1500 RAM (effective RAM = 6GHz) 7970/M = MSI 7970 Lightning 3GB, stock @ 1150/1500 RAM (forced to run at 1050/1500 when in CF to show proper scaling) 4870/A = ASUS 4870, default 750/900 core/RAM. 4870/P = Powercolor 4870, default 770/900 core/RAM, but forced to 750 core when running in CF. 4870/S = Sapphire 4870, default 750/900 core/RAM. GTX 460 = Original 460 with a 256bit memory interface. GTX 460 V2 = Later 460 with a 192bit memory interface (ironically, these overclock much better). GTX 460/E = EVGA GTX 460 (850MHz core = SSC FTW edition with 1700/1000 shader/RAM). GTX 460/P = Palit GTX 460 (800MHz core = Sonic Platinum edition with 1600/1000 shader/RAM). GTX 460/E V2 = EVGA GTX 460 V2, stock @ 867/1735/1012 GTX 280/Z = Zotac GTX 280 1GB, stock @ 700/1400/1150. GTX 580/Z = Zotac GTX 580 1.5GB, stock @ 772/1544/1002. GTX 580/E = EVGA GTX 580 1.5GB , stock @ 797/1594/1013. GTX 580/MLX = MSI GTX 580 3GB Lightning Xtreme, stock @ 832/1664/1050. GTX 750 Ti = EVGA GTX 750 Ti 2GB, stock @ 1176/1255/1350 core/boost/RAM, 344.75 drivers. GTX 980 = EVGA GTX 980 SC 4GB ACX 2.0, stock @ 1266/1367/1753 core/boost/RAM, 344.65 drivers. Quadro K5000 = PNY Quadro K5000 4GB, stock @ 706/1350 core/RAM (effective RAM = 5400MHz) Remember that the effective data rates of newer GPUs are much higher as they use newer RAM. Results are shown in order of the Avg fps. If there's a tie, then the order is by Min fps and then Max fps. ************************************************************************************************************************************************************* ********************************************************* LOW RES RESULTS *********************************************************************************** CRYSIS DX10: 1280x1024, no AA, no AF, Quality: High, NCP/CCC: Max Performance (uses ?MB VRAM): Total GPU GPU Cores / Clock VRAM Core RAM CF/ Min Avg Max CPU Threads Speed HT GPU(s) MB MHz MHz SLI FPS FPS FPS Manufac. Motherboard i7 2700K 4 / 8 5000 ON GTX 980 4096 1266 1753 NO 95.57 149.17 187.56 ASUS Maximus IV Extreme i7 2700K 4 / 8 5000 ON 2x GTX 580/MLX 3072 832 1050 YES ASUS Maximus IV Extreme i7 2700K 4 / 8 5000 ON 3x GTX 580/MLX 3072 832 1050 YES ASUS Maximus IV Extreme [SLI NVIDIA Recommended] i7 2700K 4 / 8 5000 ON 7970/S 3072 1050 1500 NO ASUS Maximus IV Extreme i7 2700K 4 / 8 5000 ON GTX 580/MLX 3072 832 1050 NO ASUS Maximus IV Extreme i5 760 4 / 4 2805 - 7970/S 3072 1050 1500 NO 48.87 77.55 108.77 Asrock P55 Extreme i5 760 4 / 4 2805 - 2x 7970/S/M 3072 1050 1500 YES 48.47 77.30 99.49 Asrock P55 Extreme i7 2700K 4 / 8 5000 ON GTX 580/Z 1536 772 1002 NO ASUS Maximus IV Extreme i3 550 2 / 4 4704 ON GTX 580/MLX 3072 832 1050 NO Asrock P55 Deluxe Pentium G3258 HW 2 / 2 3200 - GTX 750 Ti 2048 1176 1350 NO 32.84 71.73 82.64 ASUS Z97I-Plus i7 2700K 4 / 8 5000 ON Quadro K5000 4096 706 1350 NO ASUS Maximus IV Extreme i7 2700K 4 / 8 5000 ON 2x 7970/S/M 3072 1050 1500 YES ASUS Maximus IV Extreme Athlon II X2 250 2 / 2 4110 - 5850/S 1024 725 1000 NO Gigabyte GA-MA790FX-DQ6 i7 4820K 4 / 8 3900 ON GTX 460/E 1024 850 1000 NO Gigabyte GA-X79-UD3 Athlon II X2 250 2 / 2 3014 - 5850/S 1024 725 1000 NO Gigabyte GA-MA790FX-DQ6 Athlon II X2 250 2 / 2 4110 - 2x 5850/S 1024 725 1000 YES Gigabyte GA-MA790FX-DQ6 Athlon II X2 250 2 / 2 3014 - 2x 5850/S 1024 725 1000 YES Gigabyte GA-MA790FX-DQ6 i7 2700K 4 / 8 5000 ON GTX 280/Z 1024 700 1150 NO ASUS Maximus IV Extreme Athlon II X2 250 2 / 2 4110 - 4870/A 1024 750 900 NO Gigabyte GA-MA790FX-DQ6 Athlon II X2 250 2 / 2 4110 - 2x 4870/A+P 1024 750 900 YES Gigabyte GA-MA790FX-DQ6 Pentium 4 with HT 1 / 2 3400 ON 2x GTX 460/E 1024 850 1000 YES ASUS Striker II Extreme Pentium 4 with HT 1 / 2 3400 ON GTX 460/E 1024 850 1000 NO ASUS Striker II Extreme ************************************************************************************************************************************************************* ******************************************************** HIGH RES RESULTS *********************************************************************************** CRYSIS DX10: 1920x1080, 8x AA, 16x AF, Quality: Very High, NCP/CCC: Max Quality (uses MB VRAM): Total GPU GPU Cores / Clock VRAM Core RAM CF/ Min Avg Max CPU Threads Speed HT GPU(s) MB MHz MHz SLI FPS FPS FPS Manufac. Motherboard i7 2700K 4 / 8 5000 ON GTX 980 4096 1266 1753 NO 68.71 98.35 124.49 ASUS Maximus IV Extreme i7 2700K 4 / 8 5000 ON 3x GTX 580/MLX 3072 832 1050 YES ASUS Maximus IV Extreme i7 2700K 4 / 8 5000 ON 2x GTX 580/MLX 3072 832 1050 YES ASUS Maximus IV Extreme i7 2700K 4 / 8 5000 ON 2x 7970/S/M 3072 1050 1500 YES ASUS Maximus IV Extreme i7 2700K 4 / 8 5000 ON 7970/S 3072 1050 1500 NO ASUS Maximus IV Extreme i5 760 4 / 4 2805 - 2x 7970/S/M 3072 1050 1500 YES 30.57 54.34 79.79 Asrock P55 Extreme (done at 1920x1200) i5 760 4 / 4 2805 - 7970/S 3072 1050 1500 NO 23.72 44.56 56.13 Asrock P55 Extreme (done at 1920x1200) i7 2700K 4 / 8 5000 ON GTX 580/MLX 3072 832 1050 NO ASUS Maximus IV Extreme i3 550 2 / 4 4704 ON GTX 580/MLX 3072 832 1050 NO Asrock P55 Deluxe i7 2700K 4 / 8 5000 ON GTX 580/Z 1536 772 1002 NO ASUS Maximus IV Extreme Pentium G3258 HW 2 / 2 3200 - GTX 750 Ti 2048 1176 1350 NO 30.71 39.37 44.26 ASUS Z97I-Plus i7 2700K 4 / 8 5000 ON Quadro K5000 4096 706 1350 NO ASUS Maximus IV Extreme Athlon II X2 250 2 / 2 4110 - 5850/S 1024 725 1000 NO Gigabyte GA-MA790FX-DQ6 Athlon II X2 250 2 / 2 3014 - 5850/S 1024 725 1000 NO Gigabyte GA-MA790FX-DQ6 Athlon II X2 250 2 / 2 4110 - 2x 5850/S 1024 725 1000 YES Gigabyte GA-MA790FX-DQ6 i7 4820K 4 / 8 3900 ON GTX 460/E 1024 850 1000 NO Gigabyte GA-X79-UD3 Athlon II X2 250 2 / 2 3014 - 2x 5850/S 1024 725 1000 YES Gigabyte GA-MA790FX-DQ6 i7 2700K 4 / 8 5000 ON GTX 280/Z 1024 700 1150 NO ASUS Maximus IV Extreme Pentium 4 with HT 1 / 2 3400 ON GTX 460/E 1024 850 1000 NO ASUS Striker II Extreme Athlon II X2 250 2 / 2 4110 - 4870/A 1024 750 900 NO Gigabyte GA-MA790FX-DQ6 Athlon II X2 250 2 / 2 4110 - 2x 4870/A+P 1024 750 900 YES Gigabyte GA-MA790FX-DQ6 Pentium 4 with HT 1 / 2 3400 ON 2x GTX 460/E 1024 850 1000 YES ASUS Striker II Extreme ************************************************************************************************************************************************************* ********************************************************* MAX RES RESULTS *********************************************************************************** CRYSIS DX10: 2560x1600, 16x AA, 16x AF, Quality: Very High, NCP/CCC: Max Quality (uses ?MB VRAM): Total GPU GPU Cores / Clock VRAM Core RAM CF/ Min Avg Max CPU Threads Speed HT GPU(s) MB MHz MHz SLI FPS FPS FPS Manufac. Motherboard i7 2700K 4 / 8 5000 ON GTX 980 4096 1266 1753 NO 68.71 98.35 124.49 ASUS Maximus IV Extreme i7 2700K 4 / 8 5000 ON 3x GTX 580/MLX 3072 832 1050 YES ASUS Maximus IV Extreme i7 2700K 4 / 8 5000 ON 2x GTX 580/MLX 3072 832 1050 YES ASUS Maximus IV Extreme i7 2700K 4 / 8 5000 ON 7970/S 3072 1050 1500 NO ASUS Maximus IV Extreme i5 760 4 / 4 2805 - 7970/S 3072 1050 1500 NO Asrock P55 Extreme i7 2700K 4 / 8 5000 ON 2x 7970/S/M 3072 1050 1500 YES ASUS Maximus IV Extreme i5 760 4 / 4 2805 - 2x 7970/S/M 3072 1050 1500 YES . . . Asrock P55 Extreme i7 2700K 4 / 8 5000 ON GTX 580/MLX 3072 832 1050 NO ASUS Maximus IV Extreme i7 2700K 4 / 8 5000 ON Quadro K5000 4096 706 1350 NO ASUS Maximus IV Extreme Athlon II X2 250 2 / 2 4110 - 5850/S 1024 725 1000 NO Gigabyte GA-MA790FX-DQ6 Athlon II X2 250 2 / 2 4110 - 2x 5850/S 1024 725 1000 YES Gigabyte GA-MA790FX-DQ6 Athlon II X2 250 2 / 2 4110 - 4870/A 1024 750 900 NO Gigabyte GA-MA790FX-DQ6 Athlon II X2 250 2 / 2 4110 - 2x 4870/A+P 1024 750 900 YES Gigabyte GA-MA790FX-DQ6 ************************************************************************************************************************************************************* ************************************************************************************************************************************************************* Conclusions ^^^^^^^^^^^ There's clearly some kind of issue with Crossfire mode and DX9 on the Athlon II X2 250 setup, but is this to do with the hw config, an issue in the application, or perhaps the graphics drivers? The 7970 DX9/CF results suggest it's a persistent driver-related problem, as even with a fast CPU, fast GPU and the latest 14.4 drivers, DX9/CF still doesn't work properly. Thus, if you want multi-GPU for DX9, buy NVIDIA. Also, at low/res detail the lack of CPU power appears to badly limit CF scaling. Further data with other CPUs will reveal more, eg. perhaps CF really needs the presence of L3 cache to function properly? That is, as opposed to just the no. of cores or overall single-core performance. If so, then the Athlon II X4 640 should show similar behaviour to the 250 (unless more cores does help), while a Phenom II X4 965 would not; or maybe it's specific to AMD CPUs for some reason (testing with an i3 550 should be interesting). We shall see. Meanwhile, as one would expect, CPU power has less relevance at max res/detail when the GPU becomes the main bottleneck, except when the CPU is a really old single-core like the P4. As for NVIDIA cards, 3-way SLI with 580s doesn't scale properly at low resolution (2x 580 is faster). Hard to know why atm. Note that I tested with all the various available SLI modes, but it made no difference (and indeed, '3-CPU Alternate Frame Rendering 1' gave really terrible frame rates, barely 20fps). ------------------------------------------------------------------------------------------------------------------------------------------------------------ CPU-Zs: i7 2700K: http://valid.canardpc.com/show_oc.php?id=2334534 My new 6000+: http://valid.canardpc.com/2774696 Athlon II X2 250: http://valid.canardpc.com/2798885 P4/3.4/HT: http://valid.canardpc.com/z2hxcp i3 550: http://valid.canardpc.com/kzfsht John's i7 930: http://valid.canardpc.com/show_oc.php?id=1319930 John's i7 860: http://valid.canardpc.com/show_oc.php?id=1517889 My new i7 870: http://valid.canardpc.com/show_oc.php?id=1507189 i3 550: http://valid.canardpc.com/kzfsht i7 4820K: http://valid.canardpc.com/903y8q (NB: CPU-Z's RAM timings are wrong, should be 11/11/11/28/1T) i5 760 @ 2.8: http://valid.x86.fr/ly1ewa Other Benchmark Info: http://www.sgidepot.co.uk/misc/pctests.html http://www.sgidepot.co.uk/misc/tests-jj.txt http://www.sgidepot.co.uk/misc/stalkercopbench.txt http://www.sgidepot.co.uk/misc/x3tcbench.txt http://www.sgidepot.co.uk/misc/uniginebench.txt http://www.sgidepot.co.uk/misc/uniginebench2.txt http://www.sgidepot.co.uk/misc/uniginebench3.txt http://www.sgidepot.co.uk/misc/viewperf.txt http://www.sgidepot.co.uk/misc/cinebench.html http://www.sgidepot.co.uk/misc/ptboats.txt http://www.sgidepot.co.uk/misc/farcry2.txt http://www.sgidepot.co.uk/misc/callofjuarez.txt