Tom's Hardware Wants You: CPU Tests For 2011

[sp12]

I have no idea why there continues to be synthethics like 3Dmark/sisoft. There are thousands of other tech sites that will reproduce those same results to within a percent. I'd much rather see more realworld and usability tests like app-launch time with an SSD, simultaneous workloads, and maybe a source-game like L4D2.

The source games are well-known to be CPU bound, and include dozens of titles, so I feel a test with one of them would be justified.

I'd also like to see your benchmarks include at least some last-gen CPUs in order to help those on older hardware consider upgrades. For example, when SB comes out, I would expect an i7-980X, 9X0, i5-7X0, an AMD 1055t, and an AMD quad (what most tech sites use), but I would also love to see you include a Q6600/E8400 for those still kicking on LGA775.

 

[randomizer]

The problem with Handbrake is that it is far too old. The current stable release is coming up on its first birthday. AVIDemux might be a better option. It also allows you to do a lot more than Handbrake does because it's not limited to H.264 output.

[citation][nom]Cwize1[/nom]Covering the 3D rendering in the benchmarks is good but Autodesk 3ds Max 2010 is an an uber expensive piece of software. The people who use this software (legally) are the type of people who would buy whatever hardware happens to be the most expensive at the time and hence don't need these benchmarks.A more relevant tool would be blender.[/citation]

The people who use 3DS Max do so for work and probably don't pay for it because their company does. Obviously freelancers wouldn't have that luxury though. I think that if 3DS Max is going to stay in the lineup that mental ray should be used for the renderer instead of the scanline renderer though. It scales better and pushes the CPU harder. The scene should be made available for us to check out as well so we can see exactly what is being rendered. Different scenes use different features of the renderer and this affects the results.

 

[KingArcher]

Flight Simulator X

 

[dangerous_23]

some sort of HPC/scientific benchmark, not sure whats available in that regard though and I don't mean another synthetic floating point benchmark

 

[sleepflower]

I see photoshop CS5, but that really doesn't task the CPU quite like After Effects does. Most people I know use Adobe After Effects for work and would love to see a benchmark made to show which setups are best. We may not be able to download the program from you, but you could give us a folder of the project files to run our own tests on our work machines.

 

[Peciura]

Far Cry 2 or GTA4

 

[adbat]

I would sugest flash but I am not sure how to mesure it.
I would be defenetli intrested in undervolting.

 

[axipher]

Cinebench is always an option as well for testing the processor's threaded capabilities, as well as how well if performs in single threaded applications too.

 

[welshmousepk]

Id like to see some sort of multitasking benchmark. open X number of pre-determined programs to bog the system, then try tasks like opening the web browser, searching for files on the HDD, watching a video.

it may not be very quantifiable,but would be really helpful for people like me who run a million programs at once. when bulldozer and sandy bridge are around i expect multitasking to be a big deal, and less fixed by clock speed and number of cores.

 

[Guest]

Hi, you should definately consider these as well:
- Civilization V (heaviest CPU usage I have seen in a game; try zooming in & out while sliding with the arrows!)
- Starcraft 2 (one of the most popular games; lots of people interested in CPU usage running this one)

As a side note, I think synthetic test are completely useless. We want to now if we can run this & that and how, not if my core i7 is this good or not.

As another side note, I really think nowadays winrar as a benching tool is useless, most of us use 7-zip & the beta 9.xx version.

Please consider my advice & thanks anyway!

 

[Guest]

Can you guys do any Virtualization tests for cpus and gpus?

 

[Cryio]

In Physx games, you always test Physx performance from the GPU/CPU point. How about including the Fluidmark test, for the CPUs?

 

[Cryio]

Also Cryostasis (as badly optimised as it is), would be a interesting benchmark.

 

[baracubra]

I'd like to see Bad Company 2 and a source game (L4D2)
Also, GTA IV is very CPU limited so that is another good title to bench imo.

 

[Tattysnuc]

How about Superpi, and something like Superprime. Also, something like Furmark to measure the temperature.

These sorts of things are widely used, but there's no comparison for us amateur benchers vs your (pro/knowledgable) to gear our kit vs yours...

 

[Guest]

Please, add some benchmarks for programmers, for example build large projects using Visual Studio 2010 (and some Java IDE) and post the timings. CPU usage is not only about gaming, video compression, and 3D Mark

 

[lassik]

Minesweeper

 

[Guest]

Get rid of Call of Duty, Modern Warfare, every single low end config can play the game, and mid end systems with 100fps plus, pointless test!

 

[Guest]

Personally, I'd really like to see something more scientific. How about using the R-programming language (open source) to run something (or a few somethings) extremely floating-point sensitive. I'm a researcher/gamer. If anyone at the staff needs some code that takes 10 minutes/an hour or a year to run, you can find me.

 

[dominatorix]

I'd like to see Mafia II, I' like the game.

 

[lcsper]

Being an Audio Engineer, I would say Avid ProTools but I can't think of a good way to gauge performance with the software. Furthermore, most professional studios will be on ProTools|HD anyways so it doesn't really matter. Maybe loading various instances of a resource intensive virtual instrument such as Native Instrument's Kontakt or Toontracks Superior Drummer. Still, it can vary a lot. The iTunes benchmark is really inappropriate; a more resource intensive benchmark for audio should be adopted.

 

[edwilson]

Chris, one of the most usefull articles ever to hit toms is the 2 part series on bottlenecks. i have been on toms for 10 years now and work in IT myself, we have been discussing the cpu vs gpu for a LONG time. the summaries about the games that state which item impacts performance the most is really critical info. this critera testing and synopsis should be on EVERY toms article. ping me back if you need anything, be glad to do some testing for you if you need me. still running my q6600 g0

 

[Silmarunya]

Gaming test suite is primarily FPS oriented, I'd like to see some more demanding RTS titles like Empire Total War and perhaps one of the more demanding MMO titles (AoC, Aion, LOTRO or something along these lines). After all, both are very popular genres and RTS has the added advantage of being less GPU-centric than FPS games.

My second wish is that you'd run some of the programs that have both a Linux, Windows and Mac version on those other OS'es too, and the same for games (using crossover, wine and consorts). I personally use Linux for productivity and web browsing and Windows for gaming, so it'd be nice to see a comparison.

I know it'd take too long to test all apps on multiple platforms, but let's say you test 1-2 games and 1-2 productivity apps on Mac/Linux.

 

[Guest]

Supreme Commander Forged Alliance...


... over lan 2 human w/6 computers and a 1000 unit cap. It will slow down before the no rush is up. Both computers were connected on gigabit lan with quad core Phenoms and one or two dual 8800GTs (extra graphics cards made no difference)

 

[forces]

1- Crysis, 2 Crysis, 3 Crysis, 4 can be Cry... I like Resident Evil 5 too. love games that come with integrated benchmark tool...