• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Differences in power consumption between artificial and real-world heavy loads

Joseph F

Joseph F

Diamond Member
I recently wondered how much more power artificial heavy loads like Prime95 consumed compared to a real-world heavy load such as Folding@Home.
So, I decided to fire up HWiNFO64 to find out.
primeload.jpg

fahload.jpg

The top picture is Prime95 64-bit v25.11 set to in-place large FFTs.
The bottom picture was taken while running Folding@Home 7.1.43 Beta with one core dedicated to the GPU client while the remaining three were running the SMP client.

Computer specifications:
Core i5 2500k set to 4 GHz, 1.175V, turbo off, stock cooler.
16GB of G.Skill DDR3-1600.
Radeon 4850 512MB dual-slot.
Windows 7 64-bit.
 
I'm not entirely sure, but I think that it's the power that the cores alone are consuming.
It's kind of easy to forget that these new Intel chips are damn-near SOCs.
 
lol try running furmark along with your prime95. No real world test consumes that much power.

As another example check out my PC's power profile:
109d2dd360632474f532de1e6b33d0555047782070.png


First I run prime95, large FFTs. That yields 170 watts which quickly rises to 180. Then I add furmark and that is when it reaches 270.

But when I play a game like metro 2033, average power usage is only about 200-210 watts. If I run IBT instead of prime95, I think I can get the power over 300 watts, at the wall socket.
 
It's amusing that you call folding@home a "real world" workload, but not prime95.

Prime95 is just as "real world" as folding@home is. They're both distributed computing projects.
 
He's probably referring to the ' stress test ' that is part of the P95 application ..


When you are running the app to search for primes, I don't think it loads the CPU as much as the stress test does...
 
It's still doing the same exact task as it would using it for the original purpose, you're just able to change the parameters of its action a small bit, but more importantly, it doesn't try to check in and claim it's doing real work in that mode.

The use of stress testing came about because that program, by nature, was extremely taxing, the stress test wasn't put it to do different things. People were already using it as a stress tester. That functionality was added, so people would not check out work units (ok, in this case just numbers) and never check them back in because they weren't actually participating in the program. Basically, the people running the project could either add something like that, or have a ton of people hitting their servers, but doing no useful work, and in fact detracting from the overall project.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top