The Ultimate GPU Benchmark (2006 - 2010)

Published: 12th March 2022 (last update 12th March 2022 )

Two years later, part 2 is ready. New games added and of course all important video cards of selected era. This time ending with Radeon 5000 series and GeForce 400 series.

Introduction

Welcome to the second roud of ultimate GPU benchmarks. This time I focused on the DX10 era and also the first generation of the DX11 GPUs. To make full use of the test system, many multi-GPU setups are included, even 3-way SLI and CF. To add something special to this article, I decided to benchmark Nvidia Tesla C2075 card. It doesn't exactly fit the set limits of first generation Fermi, but is too interesting to leave it out. The default "Tesla mode" is very picky about drivers, so with some effort I managed to mod the BIOS to change it to Quadro 7000. Of course the GPU is still partially locked, no change here.

The test system got slight upgrade in the memory department - old DDR3 1600 was switched for DDR3 2133. Still 4x16 GB. The performance impact is almost zero... but just for the sake of completeness. :)

Test System - Hardware

Xeon E5-1650 @ 4.8 GHz (6C/6T, fixed clock)
Asus P9X79 Deluxe
4 × 4 GB DDR3 2133 CL10-12-12-31-1T
Kingston HyperX Savage 120 GB
Toshiba 3 TB / 7200 rpm
Corsair RM1000i
Fractal Design Define S
Acer FX270HU (2560×1440)

Test System - OS and Drivers

Windows 7 x64
GeForce 320.49 (for all Nvidia GPUs
Catalyst 12.1 (for Radeon HD 2000/3000/4000/5000 Series)
Catalyst 8.911.3.4000 (for FireGL V8650 OC)
...
For all GPUs forced 16×AF, high quality filtering (if possible), vsync off

Test System - Games

Bioshock [2007, DX9]
Call of Duty: World at War [2008, DX9]
Crysis [2007, DX9]
Enemy Territory: Quake Wars [2007, OGL]
F1 2012 [2012, DX10]
Far Cry 2 [2008, DX9]
GTA V [2015, DX10.1]
Mafia 2 [2010, DX9]
Metro 2033 [2010, DX10]
Mirror's Edge [2009, DX9]
Shadow Warrior [2013, DX9]
Tomb Raider [2013, DX9]

Tested Video Cards

	Radeon HD 2900 XT	FireGL V8650 OC	Radeon HD 3850	Radeon HD 3870	2 × Radeon HD 3870	3 × Radeon HD 3870	Radeon HD 4670	Radeon HD 4770
GPU	R600	R600	RV670	RV670	2 × RV670	3 × RV670	RV730	RV740
Architecture	Terascale	Terascale	Terascale	Terascale	Terascale	Terascale	Terascale	Terascale
Technology	80 nm	80 nm	55 nm	55 nm	55 nm	55 nm	55 nm	40 nm
Die Size	420 mm²	420 mm²	192 mm²	192 mm²	2 × 192 mm²	3 × 192 mm²	145 mm²	137 mm²
Transistor Count	720 mil.	720 mil.	666 mil.	666 mil.	2 × 666 mil.	3 × 666 mil.	514 mil.	826 mil.
Transistor Density	1.71 mil. / mm²	1.71 mil. / mm²	3.47 mil. / mm²	3.47 mil. / mm²	3.47 mil. / mm²	3.47 mil. / mm²	3.54 mil. / mm²	6.03 mil. / mm²
GPU Clock	743 MHz	850 MHz	670 MHz	775 MHz	775 MHz	775 MHz	750 MHz	750 MHz
Shader Clock	743 MHz	850 MHz	670 MHz	775 MHz	775 MHz	775 MHz	750 MHz	750 MHz
ROPs	16	16	16	16	2 × 16	3 × 16	8	16
TMUs	16	16	16	16	2 × 16	3 × 16	32	32
Compute Units	4	4	4	4	2 × 4	3 × 4	4	8
Shaders	320 Unified	320 Unified	320 Unified	320 Unified	2 × 320 Unified	3 × 320 Unified	320 Unified	640 Unified
L1 Cache	16 × 32 kB (Tex.)	16 × 32 kB (Tex.)	16 × 32 kB (Tex.)	16 × 32 kB (Tex.)	2 × 16 × 32 kB (Tex.)	3 × 16 × 32 kB (Tex.)	4 × 16 kB	8 × 16 kB
L2 Cache	256 kB (Tex.)	256 kB (Tex.)	256 kB (Tex.)	256 kB (Tex.)	2 × 256 kB (Tex.)	3 × 256 kB (Tex.)	128 kB	128 kB
Memory	512 MB GDDR3	2048 MB GDDR4	512 MB GDDR3	512 MB GDDR4	512 MB GDDR4	512 MB GDDR4	512 MB GDDR3	512 MB GDDR5
Memory Clock	1660 MHz	2000 MHz	1660 MHz	2250 MHz	2250 MHz	2250 MHz	2000 MHz	3600 MHz
Bus Width	512 bit	512 bit	256 bit	256 bit	256 bit	256 bit	128 bit	128 bit
Memory Bandwidth	106.2 GB/s	128 GB/s	53.1 GB/s	72 GB/s	2 × 72 GB/s	3 × 72 GB/s	32 GB/s	57.6 GB/s
Fillrate (Pixel)	11840 MP/s	13600 MP/s	10720 MP/s	12400 MP/s	2 × 12400 MP/s	3 × 12400 MP/s	6000 MP/s	12000 MP/s
Fillrate (Texel)	11840 MT/s	13600 MT/s	10720 MT/s	12400 MT/s	2 × 12400 MT/s	3 × 12400 MT/s	24000 MT/s	24000 MT/s
Compute Power (FP32)	476 GFLOPS	544 GFLOPS	428 GFLOPS	496 GFLOPS	2 × 496 GFLOPS	3 × 496 GFLOPS	480 GFLOPS	960 GFLOPS
Compute Power (FP64)	-	-	85.6 GFLOPS	99 GFLOPS	2 × 99 GFLOPS	3 × 99 GFLOPS	-	192 GFLOPS
Bus Type	PCI-E 1.0	PCI-E 1.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0
TDP	215 W	~300 W	75 W	106 W	2 × 106 W	3 × 106 W	59 W	80 W
DirectX	10	10	10.1	10.1	10.1	10.1	10.1	10.1
OpenGL	3.3	3.3	3.3	3.3	3.3	3.3	3.3	3.3
Launch Year	2007	2007	2007	2007	2007	2007	2008	2009

	Radeon HD 4830	Radeon HD 4850	Radeon HD 4850 X2	Radeon HD 4870 (512 MB) @ 625 MHz	Radeon HD 4870 (512 MB)	Radeon HD 4870 (1 GB)	Radeon HD 4890 OC	2 × Radeon HD 4890 OC
GPU	RV770	RV770	2 × RV770	RV770	RV770	RV770	RV790	2 × RV790
Architecture	Terascale	Terascale	Terascale	Terascale	Terascale	Terascale	Terascale	Terascale
Technology	55 nm	55 nm	55 nm	55 nm	55 nm	55 nm	55 nm	55 nm
Die Size	256 mm²	256 mm²	2 × 256 mm²	256 mm²	256 mm²	256 mm²	282 mm²	2 × 282 mm²
Transistor Count	956 mil.	956 mil.	2 × 956 mil.	956 mil.	956 mil.	956 mil.	959 mil.	2 × 959 mil.
Transistor Density	3.73 mil. / mm²	3.73 mil. / mm²	3.73 mil. / mm²	3.73 mil. / mm²	3.73 mil. / mm²	3.73 mil. / mm²	3.4 mil. / mm²	3.4 mil. / mm²
GPU Clock	575 MHz	625 MHz	625 MHz	625 MHz	750 MHz	750 MHz	900 MHz	900 MHz
Shader Clock	575 MHz	625 MHz	625 MHz	625 MHz	750 MHz	750 MHz	900 MHz	900 MHz
ROPs	16	16	2 × 16	16	16	16	16	2 × 16
TMUs	32	40	2 × 40	40	40	40	40	2 × 40
Compute Units	8	10	2 × 10	10	10	10	10	2 × 10
Shaders	640 Unified	800 Unified	2 × 800 Unified	800 Unified	800 Unified	800 Unified	800 Unified	2 × 800 Unified
L1 Cache	8 × 16 kB	10 × 16 kB	2 × 10 × 16 kB	10 × 16 kB	10 × 16 kB	10 × 16 kB	10 × 16 kB	2 × 10 × 16 kB
L2 Cache	256 kB	256 kB	256 kB	256 kB	256 kB	256 kB	256 kB	256 kB
Memory	512 MB GDDR3	512 MB GDDR3	1024 MB GDDR3	512 MB GDDR5	512 MB GDDR5	1024 MB GDDR5	1024 MB GDDR5	1024 MB GDDR5
Memory Clock	1800 MHz	2000 MHz	2000 MHz	3600 MHz	3600 MHz	3600 MHz	4000 MHz	4000 MHz
Bus Width	256 bit	256 bit	2 ×256 bit	256 bit	256 bit	256 bit	256 bit	2 × 256 bit
Memory Bandwidth	57.6 GB/s	64 GB/s	2 × 64 GB/s	115.2 GB/s	115.2 GB/s	115.2 GB/s	128 GB/s	2 × 128 GB/s
Fillrate (Pixel)	9200 MP/s	10000 MP/s	2 × 10000 MP/s	10000 MP/s	12000 MP/s	12000 MP/s	14400 MP/s	2 × 14400 MP/s
Fillrate (Texel)	18400 MT/s	25000 MT/s	2 × 25000 MT/s	25000 MT/s	30000 MT/s	30000 MT/s	36000 MT/s	2 × 36000 MT/s
Compute Power (FP32)	736 GFLOPS	1000 GFLOPS	2 × 1000 GFLOPS	1000 GFLOPS	1200 GFLOPS	1200 GFLOPS	1440 GFLOPS	2 × 1440 GFLOPS
Compute Power (FP64)	147 GFLOPS	200 GFLOPS	2 × 200 GFLOPS	200 GFLOPS	240 GFLOPS	240 GFLOPS	288 GFLOPS	2 × 288 GFLOPS
Bus Type	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0
TDP	95 W	110 W	250 W	150 W	150 W	150 W	190 W	2 × 190 W
DirectX	10.1	10.1	10.1	10.1	10.1	10.1	10.1	10.1
OpenGL	3.3	3.3	3.3	3.3	3.3	3.3	3.3	3.3
Launch Year	2008	2008	2008	2008	2008	2008	2009	2009

	Radeon HD 5550	Radeon HD 5670	Radeon HD 5770	Radeon HD 5770 OC	Radeon HD 5850	Radeon HD 5850 @ 850 MHz	Radeon HD 5870	Radeon HD 5870 OC
GPU	Redwood	Redwood	Juniper	Juniper	Cypress	Cypress	Cypress	Cypress
Architecture	Terascale 2	Terascale 2	Terascale 2	Terascale 2	Terascale 2	Terascale 2	Terascale 2	Terascale 2
Technology	40 nm	40 nm	40 nm	40 nm	40 nm	40 nm	40 nm	40 nm
Die Size	104 mm²	104 mm²	170 mm²	170 mm²	334 mm²	334 mm²	334 mm²	334 mm²
Transistor Count	627 mil.	627 mil.	1040 mil.	1040 mil.	2154 mil.	2154 mil.	2154 mil.	2154 mil.
Transistor Density	6.03 mil. / mm²	6.03 mil. / mm²	6.12 mil. / mm²	6.12 mil. / mm²	6.45 mil. / mm²	6.45 mil. / mm²	6.45 mil. / mm²	6.45 mil. / mm²
GPU Clock	550 MHz	775 MHz	860 MHz	1050 MHz	725 MHz	850 MHz	850 MHz	980 MHz
Shader Clock	550 MHz	775 MHz	860 MHz	1050 MHz	725 MHz	850 MHz	850 MHz	980 MHz
ROPs	8	8	16	16	32	32	32	32
TMUs	16	20	40	40	72	72	80	80
Compute Units	4	5	10	10	18	18	20	20
Shaders	320 Unified	400 Unified	800 Unified	800 Unified	1440 Unified	1440 Unified	1600 Unified	1600 Unified
L1 Cache	4 × 8 kB	5 × 8 kB	10 × 8 kB	10 × 8 kB	18 × 8 kB	18 × 8 kB	20 × 8 kB	20 × 8 kB
L2 Cache	256 kB	256 kB	256 kB	256 kB	512 kB	512 kB	512 kB	512 kB
Memory	512 MB GDDR5	1024 MB GDDR5	1024 MB GDDR5	1024 MB GDDR5	1024 MB GDDR5	1024 MB GDDR5	1024 MB GDDR5	1024 MB GDDR5
Memory Clock	4000 MHz	4000 MHz	4800 MHz	5800 MHz	4000 MHz	4800 MHz	4800 MHz	5200 MHz
Bus Width	128 bit	128 bit	128 bit	128 bit	256 bit	256 bit	256 bit	256 bit
Memory Bandwidth	64 GB/s	64 GB/s	76.8 GB/s	92.8 GB/s	128 GB/s	153.6 GB/s	153.6 GB/s	166.4 GB/s
Fillrate (Pixel)	4400 MP/s	6200 MP/s	13760 MP/s	16800 MP/s	23200 MP/s	27200 MP/s	27200 MP/s	31360 MP/s
Fillrate (Texel)	8800 MT/s	15500 MT/s	34400 MT/s	42000 MT/s	52200 MT/s	61200 MT/s	68000 MT/s	78400 MT/s
Compute Power (FP32)	352 GFLOPS	620 GFLOPS	1376 GFLOPS	1680 GFLOPS	2088 GFLOPS	2448 GFLOPS	2720 GFLOPS	3136 GFLOPS
Compute Power (FP64)	-	-	-	-	418 GFLOPS	490 GFLOPS	544 GFLOPS	627 GFLOPS
Bus Type	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0
TDP	39 W	64 W	108 W	~180 W	151 W	~180 W	188 W	~250 W
DirectX	11	11	11	11	11	11	11	11
OpenGL	4.5	4.5	4.5	4.5	4.5	4.5	4.5	4.5
Launch Year	2010	2010	2009	2009	2009	2009	2009	2009

	Radeon HD 5970	GeForce 8800 GTS (640 MB)	GeForce 8800 GTX	2 × GeForce 8800 GTX	3 × GeForce 8800 GTX	GeForce 8800 Ultra OC	Quadro FX 5600 OC	GeForce 8800 GT
GPU	2 × Cypress	G80	G80	2 × G80	3 × G80	G80	G80	G92
Architecture	Terascale 2	Tesla	Tesla	Tesla	Tesla	Tesla	Tesla	Tesla
Technology	40 nm	90 nm	90 nm	90 nm	90 nm	90 nm	90 nm	65 nm
Die Size	2 × 334 mm²	484 mm²	484 mm²	2 × 484 mm²	3 × 484 mm²	484 mm²	484 mm²	324 mm²
Transistor Count	2 × 2154 mil.	681 mil.	681 mil.	2 × 681 mil.	3 × 681 mil.	681 mil.	681 mil.	754 mil.
Transistor Density	6.45 mil. / mm²	1.41 mil. / mm²	1.41 mil. / mm²	1.41 mil. / mm²	1.41 mil. / mm²	1.41 mil. / mm²	1.41 mil. / mm²	2.33 mil. / mm²
GPU Clock	725 MHz	513 MHz	575 MHz	575 MHz	575 MHz	650 MHz	600 MHz	600 MHz
Shader Clock	725 MHz	1188 MHz	1350 MHz	1350 MHz	1350 MHz	1620 MHz	1350 MHz	1500 MHz
ROPs	2 × 32	20	24	2 × 24	3 × 24	24	24	16
TMUs	2 × 80	24	32	2 × 32	3 × 32	32	32	56
Compute Units	2 × 20	12	16	2 × 16	3 × 16	16	16	14
Shaders	2 × 1600 Unified	96 Unified	128 Unified	2 × 128 Unified	3 × 128 Unified	128 Unified	128 Unified	112 Unified
L1 Cache	2 × 20 × 8 kB	6 × 16 kB	8 × 16 kB	2 × 8 × 16 kB	3 × 8 × 16 kB	8 × 16 kB	8 × 16 kB	7 × 16 kB
L2 Cache	2 × 512 kB	96 kB	96 kB	2 × 96 kB	3 × 96 kB	96 kB	96 kB	64 kB
Memory	1024 MB GDDR5	640 MB GDDR3	768 MB GDDR3	768 MB GDDR3	768 MB GDDR3	768 MB GDDR3	1536 MB GDDR3	512 MB GDDR3
Memory Clock	4000 MHz	1600 MHz	1800 MHz	1800 MHz	1800 MHz	2200 MHz	1900 MHz	1800 MHz
Bus Width	2 × 256 bit	320 bit	384 bit	2 × 384 bit	3 × 384 bit	384 bit	384 bit	256 bit
Memory Bandwidth	256 GB/s	64 GB/s	86.4 GB/s	2 × 86.4 GB/s	3 × 86.4 GB/s	105.6 GB/s	91.2 GB/s	57.6 GB/s
Fillrate (Pixel)	2 × 23200 MP/s	10260 MP/s	13800 MP/s	2 × 13800 MP/s	3 × 13800 MP/s	15600 MP/s	14400 MP/s	9600 MP/s
Fillrate (Texel)	2 × 58000 MT/s	12312 MT/s	18400 MT/s	2 × 18400 MT/s	3 × 18400 MT/s	20800 MT/s	19200 MT/s	33600 MT/s
Compute Power (FP32)	2 × 2320 GFLOPS	228 GFLOPS	346 GFLOPS	2 × 346 GFLOPS	3 × 346 GFLOPS	415 GFLOPS	346 GFLOPS	336 GFLOPS
Compute Power (FP64)	2 × 464 GFLOPS	-	-	-	-	-	-	-
Bus Type	PCI-E 2.0	PCI-E 1.0	PCI-E 1.0	PCI-E 1.0	PCI-E 1.0	PCI-E 1.0	PCI-E 1.0	PCI-E 2.0
TDP	294 W	143 W	155 W	2 × 155 W	3 × 155 W	171 W	~171 W	125 W
DirectX	11	10	10	10	10	10	10	10
OpenGL	4.5	3.3	3.3	3.3	3.3	3.3	3.3	3.3
Launch Year	2009	2006	2006	2006	2006	2007	2007	2007

	2 × GeForce 8800 GT	GeForce 8800 GTS 512 OC	GeForce 9600 GT	GeForce GT 240	GeForce GTS 250	GeForce GTX 260 OEM	GeForce GTX 260 216	2 × GeForce GTX 260 216
GPU	2 × G92	G92	G94	GT215	G92	GT200	GT200	2 × GT200
Architecture	Tesla	Tesla	Tesla	Tesla	Tesla	Tesla	Tesla	Tesla
Technology	65 nm	65 nm	65 nm	40 nm	65 nm	65 nm	65 nm	65 nm
Die Size	2 × 324 mm²	324 mm²	240 mm²	144 mm²	324 mm²	576 mm²	576 mm²	2 × 576 mm²
Transistor Count	2 × 754 mil.	754 mil.	505 mil.	727 mil.	754 mil.	1400 mil.	1400 mil.	2 × 1400 mil.
Transistor Density	2.33 mil. / mm²	2.33 mil. / mm²	2.1 mil. / mm²	5.05 mil. / mm²	2.33 mil. / mm²	2.43 mil. / mm²	2.43 mil. / mm²	2.43 mil. / mm²
GPU Clock	600 MHz	740 MHz	650 MHz	550 MHz	740 MHz	518 MHz	576 MHz	576 MHz
Shader Clock	1500 MHz	1800 MHz	1625 MHz	1340 MHz	1836 MHz	1080 MHz	1242 MHz	1242 MHz
ROPs	2 × 16	16	16	8	16	28	28	2 × 28
TMUs	2 × 56	64	32	32	64	64	72	2 × 72
Compute Units	2 × 14	16	8	12	16	24	27	2 × 27
Shaders	2 × 112 Unified	128 Unified	64 Unified	96 Unified	128 Unified	192 Unified	216 Unified	2 × 216 Unified
L1 Cache	2 × 7 × 16 kB	8 × 16 kB	4 × 16 kB	6 × 16 kB	8 × 16 kB	8 × 24 kB	9 × 24 kB	2 × 9 × 24 kB
L2 Cache	2 × 64 kB	64 kB	64 kB	64 kB	64 kB	256 kB	256 kB	2 × 256 kB
Memory	512 MB GDDR3	512 MB GDDR3	512 MB GDDR3	512 MB GDDR5	1024 MB GDDR3	1792 MB GDDR3	896 MB GDDR3	896 MB GDDR3
Memory Clock	1800 MHz	2070 MHz	1800 MHz	3400 MHz	2000 MHz	2000 MHz	2000 MHz	2000 MHz
Bus Width	2 × 256 bit	256 bit	256 bit	128 bit	256 bit	448 bit	448 bit	2 × 448 bit
Memory Bandwidth	2 × 57.6 GB/s	66.2 GB/s	57.6 GB/s	54.4 GB/s	64 GB/s	112 GB/s	112 GB/s	2 × 112 GB/s
Fillrate (Pixel)	2 × 9600 MP/s	11840 MP/s	10400 MP/s	4400 MP/s	11840 MP/s	14504 MP/s	16128 MP/s	2 × 16128 MP/s
Fillrate (Texel)	2 × 33600 MT/s	47360 MT/s	20800 MT/s	17600 MT/s	47360 MT/s	33152 MT/s	41472 MT/s	2 × 41472 MT/s
Compute Power (FP32)	2 × 336 GFLOPS	461 GFLOPS	208 GFLOPS	257 GFLOPS	470 GFLOPS	415 GFLOPS	537 GFLOPS	2 × 537 GFLOPS
Compute Power (FP64)	-	-	-	-	-	52 GFLOPS	67 GFLOPS	2 × 67 GFLOPS
Bus Type	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0
TDP	2 × 125 W	~150 W	95 W	69 W	150 W	150 W	182 W	2 × 182 W
DirectX	10	10	10	10.1	10	10	10	10
OpenGL	3.3	3.3	3.3	3.3	3.3	3.3	3.3	3.3
Launch Year	2007	2007	2008	2009	2009	2008	2008	2008

	3 × GeForce GTX 260 216	GeForce GTX 285 OC	GeForce GTX 295	GeForce GTS 450	GeForce GTX 460 OEM	GeForce GTX 460 OC	GeForce GTX 470	GeForce GTX 480
GPU	3 × GT200	GT200B	2 × GT200B	GF106	GF104	GF104	GF100	GF100
Architecture	Tesla	Tesla	Tesla	Fermi	Fermi	Fermi	Fermi	Fermi
Technology	65 nm	55 nm	55 nm	40 nm	40 nm	40 nm	40 nm	40 nm
Die Size	3 × 576 mm²	470 mm²	2 × 470 mm²	238 mm²	332 mm²	332 mm²	526 mm²	526 mm²
Transistor Count	3 × 1400 mil.	1400 mil.	2 × 1400 mil.	1170 mil.	1950 mil.	1950 mil.	3200 mil.	3200 mil.
Transistor Density	2.43 mil. / mm²	2.98 mil. / mm²	2.98 mil. / mm²	4.92 mil. / mm²	5.87 mil. / mm²	5.87 mil. / mm²	6.08 mil. / mm²	6.08 mil. / mm²
GPU Clock	576 MHz	700 MHz	576 MHz	810 MHz	650 MHz	925 MHz	608 MHz	700 MHz
Shader Clock	1242 MHz	1476 MHz	1240 MHz	1620 MHz	1300 MHz	1850 MHz	1216 MHz	1400 MHz
ROPs	3 × 28	32	28	16	32	32	40	48
TMUs	3 × 72	80	80	32	56	56	56	60
Compute Units	3 × 27	30	2 × 30	4	7	7	14	15
Shaders	3 × 216 Unified	240 Unified	2 × 240 Unified	192 Unified	336 Unified	336 Unified	448 Unified	480 Unified
L1 Cache	3 × 9 × 24 kB	10 × 24 kB	2 × 10 × 24 kB	4 × 64 kB	7 × 64 kB	7 × 64 kB	14 × 64 kB	15 × 64 kB
L2 Cache	3 × 256 kB	256 kB	2 × 256 kB	256 kB	512 kB	512 kB	768 kB	768 kB
Memory	896 MB GDDR3	1024 MB GDDR3	896 MB GDDR3	1024 MB GDDR5	2048 MB GDDR5	1024 MB GDDR5	1280 MB GDDR5	1536 MB GDDR5
Memory Clock	2000 MHz	2800 MHz	2000 MHz	3600 MHz	3400 MHz	4200 MHz	3350 MHz	3700 MHz
Bus Width	3 × 448 bit	512 bit	2 × 448 bit	128 bit	256 bit	256 bit	320 bit	384 bit
Memory Bandwidth	3 × 112 GB/s	179.2 GB/s	2 × 112 GB/s	57.6 GB/s	108.8 GB/s	134.4 GB/s	134 GB/s	177.6 GB/s
Fillrate (Pixel)	3 × 16128 MP/s	22400 MP/s	2 × 16128 MP/s	12960 MP/s	20800 MP/s	29600 MP/s	24320 MP/s	33600 MP/s
Fillrate (Texel)	3 × 41472 MT/s	56000 MT/s	2 × 46080 MT/s	25920 MT/s	36400 MT/s	51800 MT/s	34048 MT/s	42000 MT/s
Compute Power (FP32)	3 × 537 GFLOPS	708 GFLOPS	2 × 595 GFLOPS	622 GFLOPS	874 GFLOPS	1243 GFLOPS	1089 GFLOPS	1344 GFLOPS
Compute Power (FP64)	3 × 67 GFLOPS	88.5 GFLOPS	2 × 74 GFLOPS	52 GFLOPS	73 GFLOPS	104 GFLOPS	136 GFLOPS	168 GFLOPS
Bus Type	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0
TDP	3 × 182 W	~225 W	289 W	106 W	150 W	~225 W	215 W	250 W
DirectX	10	10	10	11	11	11	11	11
OpenGL	3.3	3.3	3.3	4.6	4.6	4.6	4.6	4.6
Launch Year	2008	2009	2009	2010	2010	2010	2010	2010

	GeForce GTX 480 OC	Quadro 4000	Tesla C2075 Mod
GPU	GF100	GF100	GF110
Architecture	Fermi	Fermi	Fermi
Technology	40 nm	40 nm	40 nm
Die Size	526 mm²	526 mm²	520 mm²
Transistor Count	3200 mil.	3200 mil.	3000 mil.
Transistor Density	6.08 mil. / mm²	6.08 mil. / mm²	5.77 mil. / mm²
GPU Clock	825 MHz	475 MHz	700 MHz
Shader Clock	1650 MHz	950 MHz	1400 MHz
ROPs	48	32	48
TMUs	60	32	56
Compute Units	15	8	14
Shaders	480 Unified	256 Unified	448 Unified
L1 Cache	15 × 64 kB	8 × 64 kB	14 × 64 kB
L2 Cache	768 kB	768 kB	768 kB
Memory	1536 MB GDDR5	2048 MB GDDR5	6144 MB GDDR5
Memory Clock	4000 MHz	2800 MHz	3700 MHz
Bus Width	384 bit	256 bit	384 bit
Memory Bandwidth	192 GB/s	89.6 GB/s	177.6 GB/s
Fillrate (Pixel)	39600 MP/s	15200 MP/s	33600 MP/s
Fillrate (Texel)	49500 MT/s	15200 MT/s	39200 MT/s
Compute Power (FP32)	1584 GFLOPS	486 GFLOPS	1254 GFLOPS
Compute Power (FP64)	198 GFLOPS	243 GFLOPS	627 GFLOPS
Bus Type	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0
TDP	~350 W	142 W	225 W
DirectX	11	11	11
OpenGL	4.6	4.6	4.6
Launch Year	2010	2010	2011