What are your favorite benchmarking apps?

[zCypher]

I'd like to stick to free apps if possible!

I have Super PI / mod1.5 XS (not sure of the differences between various versions really).
I always run Windows Experience Index, though I know it's not considered to be all too meaningful.

When tinkering with overclocking, I like to keep HWMonitor + CPU-Z open to keep an eye on things, but this doesn't give me much benchmarking info aside from temperatures/voltages at a given clockrate.

ATTO and CDM give me a good idea of what's happening when it comes to hard drives, but what about CPU and RAM? Any good free ones out there that give reliable data?

 

[denev2004]

Sisoftware Sandra Lite Free...It's really professional.

 

[Blastman]

This Fritz 12 chess benchmark uses as many cores or threads as available.

Other free benches you can google include …

Cinebench 11.5
wPrime 32M

 

[sm625]

My favorite benchmark is a batch script that opens 4 web browsers with 3 tabs each, 5 adobe pdf documents, a jpg in windows fax viewer, windows live mail, a word doc, and xls spreadsheet, an openoffice doc, an openoffice spreadsheet, a powerpoint doc in powerpoint viewer, a txt file in notepad, a few explorer windows.

Before running the script I start one prime95 thread.

One of the pdf files is huge. While all this stuff is loading I try to scroll through that document and note the smoothness of it.

 

[Chiropteran]

I run a bitcoin miner, and note the mhash value and power usage of the computer.

Alternatively, I play a computer game and note my personal level of enjoyment.

The only benchmarks I really care about are real world performance

 

[Ferzerp]

^^^ because saying 7zip would have been too obvious

 

[Golgatha]

I only use benchmark apps to determine stability. To that end I like OCCT, Prime95, 3D Mark 11, and Furmark. If a system can survive running each of those for a few hours each and running 3D Mark looping overnight, I would consider the system stable.

I run a few passes of Memtest x86 and boot Knoppix from a live CD before a Win7 install as well, just to make sure the hardware is all working before using any of these benchmarks. I tend to be very, very demanding on my hardware, so I burn it in very harshly to make sure it can handle my daily abuse (24/7 Folding@Home, video encoding/transcoding, Virtual machines, heavy gaming, etc).

 

[fffblackmage]

I usually use Prime 95 to test for stability. I don't even run it long - no more than 1 hour. For actual benchmarking, I like to run game benchmarks (like STALKER:COP, FC2), since that's really the only cpu-intensive thing I do.

 

[zCypher]

Sweet, thanks for all the suggestions, grabbing them now.
Wow 3dmark gives me some very bad framerates lol http://3dmark.com/3dm11/2780732, furmark gives me 9fps :|

in cinebench i scored 49.13 in the opengl test, 6.86pts in CPU test, 1.16 in single core CPU test, and 5.89x MP ratio.

 

[exar333]

^^^ because saying 7zip would have been too obvious

LOL!!

 

[Chiropteran]

^^^ because saying 7zip would have been too obvious

I get it. Very clever.

I guess I could also say fps/dollar, there is another metric where the CPUs I use win

 

[exar333]

This Fritz 12 chess benchmark uses as many cores or threads as available.

Other free benches you can google include …

Cinebench 11.5
wPrime 32M

I like those too; added a few as well:

SiSoft Sandra (free!)
3DMark 11
OCCT (Stability)
Linpack (Stability)
Memtest (Stability)

CPU-ID (CPU-Z) is also really nice when helping other folks. You can download it and immediately know the CPU, RAM, and GPU details without even cracking it open. You know the # of DIMM slots, chipset, etc. Its really a time-saver.

 

[exar333]

I get it. Very clever.

I guess I could also say fps/dollar, there is another metric where the CPUs I use win

(Putting on my devil's advocate hat here)

You have a $60 G630 that gets the same fps when gaming as a $269 8150.

I take it you own the G630?

 

[Chiropteran]

Almost.

It's a cheap Athlon X2 bought for mining, but when running a GPU limited game it does just fine.

 

[Ferzerp]

I get it. Very clever.

I guess I could also say fps/dollar, there is another metric where the CPUs I use win

I thought we are all past the point where we try to pass off GPU limited benchmarks as valid comparisons of CPUs. It's such a redumentary mistake to make that I would imagine no one would make it anymore unless it is on purpose. It's as useful as benchmarking CPUs by measuring the throughput with a given network card.

Perhaps some people are more interested in promoting something than actually getting an accurate measure of useful information.

 

[Chiropteran]

Perhaps some people are more interested in promoting something than actually getting an accurate measure of useful information.

Ah. You are right.

It's much more important to know how fast your CPU can complete a completely artificial task which is not used by any software you ever touch in day to day use, rather than measuring it's performance in software you do actually use.

edit: Also, please activate your sarcasm detector. I was responding tongue in cheek to the guy saying 7-zip was an important benchmark.

 

[Ferzerp]

I think you mean measuring your GPUs performance on software you use.

Yes, it is more useful when comparing CPUs to measure them artificially than to measure your GPU and claim that it measures your CPU.

 

[Chiropteran]

I think you mean measuring your GPUs performance on software you use.

Yes, it is more useful when comparing CPUs to measure them artificially than to measure your GPU and claim that it measures your CPU.

Last time I looked, there are performance difference between using (for example) an i7 2700k vs an FX-4100. Sometimes the difference is even noticeable! Obviously you aren't measuring the GPU when you use the exact same GPU on both systems and get differing results.

Let me make up a hypothetical situation. CPU A performs at X fps in all games you play. CPU B performs at (X + 10%) fps in all games you play. However, CPU A can complete artificial benchmarks A, B, and C an incredible 50% faster rate than CPU B.

What you are arguing right now is that you would buy CPU A due to it's advantage in artificial benchmarks, even though it is inferior at running the actual applications (games) you run.

You are irrational, IMO.


However, my (sarcastic) point was that when you take the difference in FPS and compare with the difference in dollars spent, the advantage falls to the cheaper option.

 

[nemesismk2]

I like 3dmark 11

 

[Ferzerp]

I guess I was wrong when I said no one would fall for the mistake of measuring one component's capability by directly measuring the capability of another component.

Unless you do that on purpose to try to obfuscate something.

We're done here, as I have read enough of your posts to know that when the silly bat avatar appears, truth is not important. What is are words that paint AMD in the best possible light in all situations.

 

[Chiropteran]

What is are words that paint AMD in the best possible light in all situations.

I'm not even sure that was English, but I think I get what you are trying to say.

However, I'd suggest you look at my original post and please try to tell me how it was pushing AMD CPU in any way. Intel CPU can mine bitcoins fine (given a decent AMD video card), and they should be capable of running the games I play without serious issues. If they didn't cost so much more I'd be building my mining rigs with Intel boards.

It was YOU who decided to turn this into an argument by making your snarky comment about 7-zip. I guess you like to dish it out but you can't handle it when you get a response in kind.

 

[nyker96]

I don't trust any one program for cpu benchmarking since the result of one can hardly give you the entire picture. I just do this. If I'm interested in how fast it does in games, I just use specific games like Crysis etc. If I'm interested in video compression speed, I use x264 benchmarks. if data compression I use 7zip etc etc. you get the picture.

 

[BTRY B 529th FA BN]

If your data is not mission critical then do maybe an hour or two of LinX or P95 and be done with it and after that game on, if you are overclocking for better gaming performance. No use running your pc all night just to have it crash for some other reason than stability like a bogus driver. IMO stability testing is a waste of time if data isn't mission critical and if it is, don't overclock your pc, and if you do keep it reasonable. I don't keep my main rig at any overclock just stock speed.

 

[zCypher]

Is Linpack and LinX the same thing?
What's an example of a x264 benchmark?

 

[wildside50]

It's KIND of a proprietary benchmark, but it's the DOESITPISSMEOFFMARK. Recently my Athlon 3200+ failed it because it dropped frames on a blu ray rip. Then my Opteron 180 failed it because it rebooted randomely. My G620 just passed it with flying colors.