PSU (power supply unit) wears out in time. It can fail earlier if it's not properly sized (loaded close to max output watts) or if the electric network provider doesn't provide stabilized current (variations in voltage)...
I have little time today. I have an old Quadro on my desk to test it, I'll test it and share results a.s.a.p., hopefully in about 6 hours.
The first link I provided, in the first post, which is huge, made by bambiboom, you'll find various results with T5400. Just click see full contents under Mates, it doesn't show in the first place because it's huge.
Code:
***************************************
Content of this section is hidden, You must be registered and activate your account to see this content. See this link to read how you can remove this limitation:
http://forum.civilea.com/thread-27464.html
***************************************
I found this benchmark, I will provide results with it, SPECviewperf 11:
Code:
***************************************
Content of this section is hidden, You must be registered and activate your account to see this content. See this link to read how you can remove this limitation:
http://forum.civilea.com/thread-27464.html
***************************************
I'm using this setup for a while ~2 years:
Kingston 8GB DDR3 1333MHz CL9 Dual Channel Kit Non ECC
WD 500 GB SATA-III 32MB 7200rpm Caviar Black
Intel Core i5 2500 3.30GHz
MaxCube Nira 500W
Asus P8H61-M LE
DVD-Writer Optiarc AD-7280S
Inter-Tech Hedgehog
It's price as a new system was about 486 US$.
The power consumption is about 95W CPU, 20W Hdd, 50W Mainboard+ram+usb~170W
Most of the time less because it's not loaded at max all the time.
As you can see no GPU, I'm using the embedded graphics card which is free and doesn't drain hundreds of watts.
Results with SpecviewPerf, embedded Intel 2000 HD:
Quote:Viewset Composite Multisample Performance
catia-03 not Successful
ensight-04 1.85 no result
lightwave-01 11.76 no result
maya-03 10.90 no result
proe-05 0.79 no result
sw-02 5.39 no result
tcvis-02 1.27 no result
Some hardware expensive results:
Code:
***************************************
Content of this section is hidden, You must be registered and activate your account to see this content. See this link to read how you can remove this limitation:
http://forum.civilea.com/thread-27464.html
***************************************
Comparison:
Code:
***************************************
Content of this section is hidden, You must be registered and activate your account to see this content. See this link to read how you can remove this limitation:
http://forum.civilea.com/thread-27464.html
***************************************
My results are close to a GTX 680
. Which is absurd. Moreover there's a huge difference between my cpu and the tested one, see here:
Code:
***************************************
Content of this section is hidden, You must be registered and activate your account to see this content. See this link to read how you can remove this limitation:
http://forum.civilea.com/thread-27464.html
***************************************
1200 W PSU
You may find here the cpu results, I can confirm those values, and those are pretty close to what T5400 offers:
Code:
***************************************
Content of this section is hidden, You must be registered and activate your account to see this content. See this link to read how you can remove this limitation:
http://forum.civilea.com/thread-27464.html
***************************************
On the other hand look at single core results, those really matter. Most of the stuff runs most of the time in single core, including SpecviewPerf, sap, etabs, tekla, autocad, and whatever, except for ansys which scales very well on multiple cores.
T5400 offers 1616 while i5 offers 3000, this is from single to double.
The difference is because of different technology in cpu:
Code:
***************************************
Content of this section is hidden, You must be registered and activate your account to see this content. See this link to read how you can remove this limitation:
http://forum.civilea.com/thread-27464.html
***************************************
To check if your favorite software uses multithreading simply start task manager and look at processes, the cpu tab, if it's 100% and you have multicore it does, if it's 25% and you have 4 cores it doesn't and you need high perf in single core mode.
The AMD FX-8350 really got my attention the 8 cores cpu. I wish I had the opportunity to test it with real stuff like solvers or OpenGL, not games.
I also use a Kingston 60GB SATA-III 2.5 inch V300 SSDNow, 65 US$ for storing analysis results or performing large structure nonlinear analysis. (The sap/etabs null steps take like 100 times less time to perform). Viewing results takes like double less time to perform when storing data on it. It's very very cheap for the huge performance it provides. (compared to other cpu price or ram mhz, or overclocking).
To calculate power requirements (google for power requirements/supply calculator):
Code:
***************************************
Content of this section is hidden, You must be registered and activate your account to see this content. See this link to read how you can remove this limitation:
http://forum.civilea.com/thread-27464.html
***************************************
I can perform analysis in sap for a specific model and provide the time spent with cpus liek q9550, i5-2500, core 2 duo, i5-3550. In about a week I can do it also with i7-4771.
I can say that especially IntelBurnTest v2.54 and maybe Nuclearus Multicore v2.00 will give close results with the real world results.
I mean by that in IntelBurnTest I see 85.3 GFlops and Time of 31 s for stress high, with i5-2500 and I see about 30 GFlops for q9550 and the same sap model runs faster with almost same ratio.