GeForce GTX 960

GPU-accelerated LIBSVM を動かしたくてGeForce GTX 960を購入。

下記、CUDA-Zの結果。

CUDA-Z Report

Version: 0.9.231 http://cuda-z.sf.net/
OS Version: Windows AMD64 6.1.7601 Service Pack 1
Driver Version: 347.25
Driver Dll Version: 7.0 (8.17.13.4725)
Runtime Dll Version: 6.0

Core Information

Name	GeForce GTX 960
Compute Capability	5.2
Clock Rate	1228 MHz
PCI Location	0:1:0
Multiprocessors	8
Therds Per Multiproc.	2048
Warp Size	32
Regs Per Block	65536
Threads Per Block	1024
Threads Dimensions	1024 x 1024 x 64
Grid Dimensions	2147483647 x 65535 x 65535
Watchdog Enabled	Yes
Integrated GPU	No
Concurrent Kernels	Yes
Compute Mode	Default
Stream Priorities	No

Memory Information

Total Global	2048 MiB
Bus Width	128 bits
Clock Rate	3505 MHz
Error Correction	No
L2 Cache Size	48 KiB
Shared Per Block	48 KiB
Pitch	2048 MiB
Total Constant	64 KiB
Texture Alignment	512 B
Texture 1D Size	65536
Texture 2D Size	65536 x 65536
Texture 3D Size	4096 x 4096 x 4096
GPU Overlap	Yes
Map Host Memory	Yes
Unified Addressing	No
Async Engine	Yes, Bidirectional

Performance Information

Memory Copy
Host Pinned to Device	2508.83 MiB/s
Host Pageable to Device	1433.22 MiB/s
Device to Host Pinned	1552.14 MiB/s
Device to Host Pageable	1034.73 MiB/s
Device to Device	38.3993 GiB/s
GPU Core Performance
Single-precision Float	2712.13 Gflop/s
Double-precision Float	86.5304 Gflop/s
32-bit Integer	821.532 Giop/s
24-bit Integer	602.154 Giop/s

Generated: Sat Feb 07 12:09:00 2015




960を挿したベースのPCがPCIe3に対応していないからか目的のLibSVMはノートPCのCPUの方が速いという残念な結果でした。