Built a very similar system a while ago as an home server
Same MB
E3-1245v2 CPU
same memory
same drives 24*3TB WD Red drives
Raid controller is the new Adaptec 7 Series with Flash memory based backup system
Energy consumption is ca. 150 watt on the wall outlet when all disk spindles are on...
Experienced significant sample variation with Titan cards, very low sample variation with 7970 cards.
Asus GTX Titan, stock clock: (1920x1080 in windowed mode, Ultra&Extreme setting)
cards were set with EVGA precision to default mode. Fan was set to 75%
Titan_1: 1400 pts (actual speed was 1005...
This is to be seen production code, as this is currently in beta
If you take the FahCore17 Benchmark utility build on the Open CL interface, the GTX Titan is slightly faster that the 7970. (With Cuda, the Titan is 58% faster than with its OpenCL interface). I recently added 4 x 7970 and like in...
No :)
Two systems:
1) single i7-3930K/X79 board (picture above)
2) dual Xeon E5-2665 board with better data transfer rates memory / GPUs (currently being built)
switching GPUs back an forth to check implications and performance on different architectures/software combinations. Mostly...
I did run all 3 LuxMark 2.0 on both platforms (check out user AndreasE in the LuxMark database links below)
LuxMark 2.0 room: http://www.luxrender.net/luxmark/top/top20/Room/GPU
4 x AMD 7970: 5317 (currently #3)
4 x GTX Titan: 2788 (= 52%)
LuxMark 2.0 sala...
Agree that not all programs are stream benchmarks. My comment was a response to the statement above that basically most perf increase comes from the "core" CPU.
BTW, many HPC systems turn SMT off - the ability of hiding latency in this class of apps is limited. Turning on SMT quite often leads...
Don't get caught by the increase in bandwidth. The primary bottleneck today is latency (access time). This access time (measured in nanoseconds) increased since the IBM PC days only 4-5 fold. That's the primary reason why we got one level of cache, then a second and now a third.
These days...
Don't forget memory hierarchy scaling.
Over time, if memory requirements keep at a constant factor, the application's memory footprint "moves up" the cache hierarchy.
Just for fun did I use a virtual machine with DOS 5 and Turbo Pascal. My 30.000 line chess program from the 90ties...
Yes, I know :)
I changed the settings with EVGA precision to:
shader +160 MHz
memory + 160 MHz
Power = 100% (default)
Temp = 80 C (default)
Fans = 85% (max)
While the fans are noisy at this settings the cards keep running at ca. 70 degree C (Usually, I keep the system loaded for 24 hrs...
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.