KFA2 GeForce GTX 980 ‘8Pack Edition’ 4096MB

1. Introduction2. KFA2 GeForce GTX 980 '8Pack Edition'3. High Res Gallery4. Testing Methodology5. 3DMark Vantage6. 3DMark 117. 3DMark8. Unigine Heaven Benchmark9. Grid AutoSport (1080p and 1600p)10. Grid AutoSport (4k)11. Thief (1080p and 1600p)12. Thief (4k)13. Tomb Raider (1080p and 1600p)14. Tomb Raider (4k)15. Metro Last Light Redux (1080p and 1600p)16. Metro Last Light Redux (4k)17. Thermal Dynamics / IR Thermometer Readings18. Acoustics Performance19. Power Consumption20. Overclocking21. Closing Thoughts22. View All Pages

Nvidia's GTX980 has been one of their greatest success stories in the last decade. While we have reviewed almost every partner board available, today we look at something very special – the new KFA2 GeForce GTX980 ‘8Pack Edition'. Overclockers UK resident overclocking world champion Ian ‘8Pack' Parry has been working with the company to produce one of the world's fastest air cooled cards. This solution is not only shipped at very high overclocked speeds, but it has been the recipient of uprated Capacitors and VRM components.

Nvidia's Maxwell architecture has been dominating the graphics sector for the last year. The GTX750ti was a warning shot across the bow for AMD – a super efficient GPU with more powerful siblings waiting in the wings. Some time later, the GTX970 and GTX980 hit market without an answer. AMD's power consumption is still a concern and until their next generation of cards hit later this year Nvidia have been able to run amok in the high end.

GPU	GeForce GTX 750ti (Maxwell)	GeForce GTX 960 (Maxwell)	Geforce GTX 970 (Maxwell)	GeForce GTX 980 (Maxwell)	Geforce GTX Titan X
Streaming Multiprocessors	5	8	13	16	24
CUDA Cores	640	1024	1664	2048	3072
Base Clock	1020 mhz	1126 mhz	1050 mhz	1126 mhz	1000 mhz
GPU Boost Clock	1085 mhz	1178 mhz	1178 mhz	1216 mhz	1076mhz
Total Video memory	2GB	2GB	4GB	4GB	12GB
Texture fill-rate	40.8 Gigatexels/Sec	72.1 Gigatexels/Sec	109.2 Gigatexels/Sec	144.1 Gigatexels/Sec	192 Gigatexels/Sec
Memory Clock	5400 mhz	7010 mhz	7000 mhz	7000 mhz	7000 mhz
Memory Bandwidth	86.4 GB/sec	112.16 GB/sec	224 GB/s	224 GB/sec	336.5 GB/sec
Bus Width	128bit	128bit	256bit	256bit	384bit
ROPs	16	32	56	64	96
Manufacturing Process	28nm	28nm	28nm	28nm	28nm
TDP	60 watts	120 watts	145 watts	165 watts	250 watts

The new GM204 GPU is very efficient. The Maxwell SM has been rebalanced so that the CUDA cores are fully utilised more often. Doing so saves power and enhances overall performance. The L2 cache size in the GM204 is 2MB, or four times larger than the GK104. The addition of extra cache means that fewer requests to the GPU memory are needed – again reducing power consumption and pushing more performance.

The Geforce GTX 980 is equipped with 7Gbps memory. Those of you with higher resolution monitors, or running in a multi screen configuration will also be pleased to hear that there is 4GB of GDDR5 memory on the card, not 3GB – this memory will push 224 GB/s sec. Nvidia have added a new compression engine to reduce the demand on DRAM bandwidth.

The KFA2 Geforce GTX980 ‘8Pack Edition' has received a mighty impressive core clock enhancement to 1,304mhz – from 1,126mhz on the reference design (boost speed 1418mhz). The GDDR5 memory is running at the default 1,753mhz speeds (7Gbps effective) but we believe there will be huge potential for overclocking, later in the review.

Become a Patron!