aglarond
Distinguished
- Sep 30, 2010
- 11
- 0
- 18,510
- Status
[citation][nom]pakardbell486dx2[/nom]Even though this tread is for software, I still think tom's should list single core cpu's like Pentium 4's and AMD athlon 64's into the tests, becasue a large portion of us still use those old platforms. And even use cards based on agp format as well. I know those options are long dead and are no longer selling but like i said a large amount of us are still using that becasue a) they work fine and b) don't have the money for a new setup.[/citation]
This could be helpful, for someone with an older system to gauge whether they can even play that shiny new game that just came out...unfortunatly, I think it would clog up the system too much. Maybe as a seperate page? At most, 2-3 "old" CPUs ought to be included in the main chart, and users can make educated guesses based on those numbers.
It's def good to support the old stuff, but it would be annoying if the chart became cluttered with many, ancient, CPUs. IMHO 😀
This could be helpful, for someone with an older system to gauge whether they can even play that shiny new game that just came out...unfortunatly, I think it would clog up the system too much. Maybe as a seperate page? At most, 2-3 "old" CPUs ought to be included in the main chart, and users can make educated guesses based on those numbers.
It's def good to support the old stuff, but it would be annoying if the chart became cluttered with many, ancient, CPUs. IMHO 😀
MrPerlishells
Distinguished
- Sep 19, 2008
- 15
- 0
- 18,510
Ruse. It has its own benchmark that test CPU and GPU together and independently.
Blender 2.49b or 2.50 when it is released. Its an open source 3D program that just recently released a short called Sintel. I believe this has bee used before.
Distributed.net RC5-72 cracking challenge. This has been going on for awhile now. Until recently has slowed to a crawl taking years to complete less than 1% of the keyspace. Recent addition allows the use of GPU both ATI and nVidia to crack codes. RC5 is an encryption algorithm developed by RSA used in Internet Explorer.
Blender 2.49b or 2.50 when it is released. Its an open source 3D program that just recently released a short called Sintel. I believe this has bee used before.
Distributed.net RC5-72 cracking challenge. This has been going on for awhile now. Until recently has slowed to a crawl taking years to complete less than 1% of the keyspace. Recent addition allows the use of GPU both ATI and nVidia to crack codes. RC5 is an encryption algorithm developed by RSA used in Internet Explorer.
How about something testing the CPU running CUDA code? This would be more a benchmark of the GPU, and then the motherboard (PCIe transfer speeds.) There are enough math geeks and folding farm people out there that this benchmark would be appreciated by some.
[citation][nom]RySean[/nom]Personally, as old as it is, I'd still like to see Crysis show up on benchmarks. Although, it's fairly GPU limited, so it may be better suited for GPU benchmarks.It's still one of the more challenging games to run on even a modern rig...[/citation]
Agreed...Crysis still needs to be in there since it can still bring many systems to their knees with full eye candy on. MW2 doesn't even bring my hardware temps up from idle so not sure why it is in there. Dirt 2 is a good one.
Agreed...Crysis still needs to be in there since it can still bring many systems to their knees with full eye candy on. MW2 doesn't even bring my hardware temps up from idle so not sure why it is in there. Dirt 2 is a good one.
Chris, lose the Apple iTunes 10.0.1 anybody on top of the game has switched to flak or still uses audiocatalyst because it gives out the highest cd quality encoding you can get.
add in a bench mark showing multiplayer lan game and or lan file transfer, voice chat client&or w/wo instant messenger, a couple online webpages, downloading a file, playing some music/music videos and burning/copying a cd/dvd all being done at the same time. this is something i know me and my friends have been doing since 1997 and are still doing today. i understand the difficulty in this, but hey it's a real world simulation (plus it means you guys will actually get to play a game while benching, got to be worth it there, right?)
this will really push the systems and i hope, show what a multicore cpu is capable of in real world worthiness.
add in a bench mark showing multiplayer lan game and or lan file transfer, voice chat client&or w/wo instant messenger, a couple online webpages, downloading a file, playing some music/music videos and burning/copying a cd/dvd all being done at the same time. this is something i know me and my friends have been doing since 1997 and are still doing today. i understand the difficulty in this, but hey it's a real world simulation (plus it means you guys will actually get to play a game while benching, got to be worth it there, right?)
this will really push the systems and i hope, show what a multicore cpu is capable of in real world worthiness.
Bad company 2 and Civilisation V as probably the most cpu intensive games atm, and Bad company 2 being one of the few that perform significantly better on an amd 6core than a 4core i7
thechief73
Distinguished
- Feb 8, 2010
- 1,126
- 0
- 19,460
First I would say Far Cry 2, but the requirements of a widely used or adopted product is most likly not met with this title. I loved the graphics and the enviroment was awe inspiring even if there were some minor annoyances to the gameplay, I think its one of those great ones that slipped through the gamers radar.
Ok, I know this probably has been said 100 times now (Heck its in the second post, lol) but really I cannot think of any game as of now that I would rather see benches on more than Crysis. And by mid 2011 you can add Crysis 2 to the mix, please =)
Ok, I know this probably has been said 100 times now (Heck its in the second post, lol) but really I cannot think of any game as of now that I would rather see benches on more than Crysis. And by mid 2011 you can add Crysis 2 to the mix, please =)
I'm probably going to get stoned and crucified for saying this but World of Warcraft is getting many graphical updates with its next expansion. Its always been a CPU heavy game it also supports multi cpu core and multi gpu setups quite well. And despite being CPU heavy it has up to date supported SLI and Crossfire quite well. Can we see some benchmarks of WoW post-cataclysm graphical changes?
pjmelect
Splendid
- Jul 14, 2006
- 3,216
- 23
- 21,415
I would like to see a single threaded benchmark so that I can see the performance of a single core with results that give an easily comparable number of the performance against other processors. There is a benchmark that does this that is called Metabench available from 7byte at http://7byte.com/index.php?page=metabench
This benchmark is a bit buggy and long in the tooth but is the benchmark I use when comparing processors. I can then do a mental adjustment with the results to take into account the number of cores.
This benchmark is a bit buggy and long in the tooth but is the benchmark I use when comparing processors. I can then do a mental adjustment with the results to take into account the number of cores.
BadKharmaCDN
Distinguished
- Apr 29, 2008
- 31
- 0
- 18,530
try adding seti . it may not be completely accurate on a per packet time , but over 20 or 30 packets averaged might give some idea about cpu horse power
PassMark
Distinguished
- Jan 19, 2004
- 428
- 35
- 19,040
Although I have to declare some self interest in this, synthetic CPU benchmarks are useful becuase,
A) They can generally be automated and in most cases can run fairly quickly. So the additional overhead for Tom's of including the synthetics in an automated suite is low.
B) They pretty much all have a free trial, or are totally free.
C) They are (for the most part) small downloads with a easy install and run a standardised test.
D) Users reading Tom's site can compare their home / work systems with Tom's results. If you are testing with Office suites, 3D suites and even games it can be difficult to reproduce the test results as you won't have the same set of test files & scripts (unless Tom makes all this available).
E)Patches to games and other applications can effect performance. WoW now using all cores being a good example. So getting and keeping the exact version of an app becomes critical for a fair comparision. In many cases old releases of commerical applications are not available to the public. Meaning results can not be reproduced by the public once some time as passed. The synthetics are more careful that small patches don't effect performance and generally make old releases available.
E) Some of the above suggestions would not make good CPU benchmarks by virtue of the fact that they are not CPU bound. Microsoft Security Essentials, Office applications and even Visual Studio can often be disk bound rather than CPU bound. Some of the other suggestions like Photoshop, Flash, SETI, etc might also be impacted by the 2D video performance of the video card or networking performance. The synthetic CPU benchmarks don't generally have much if any depenance on the video card & disk.
F) The synthetics can in general use all cores (or a configurable number of cores). By the end of 2011 it will be common place for CPUs to have 8 to 16 cores (with HT). So it is important to have at least some of the benchmarks pushing all core to their limit. What current games are going to use 100% of 16 cores?
G) The synthetics are actually not all that synthetic in many cases (ignoring the simple integer and floating point tests). They often use real world algorithms pulled from real world applications.
So I would be in favour of have some synthetics in the list.
PCMark, SiSoftware, PerformanceTest CPUMark, SuperPi, etc..
A) They can generally be automated and in most cases can run fairly quickly. So the additional overhead for Tom's of including the synthetics in an automated suite is low.
B) They pretty much all have a free trial, or are totally free.
C) They are (for the most part) small downloads with a easy install and run a standardised test.
D) Users reading Tom's site can compare their home / work systems with Tom's results. If you are testing with Office suites, 3D suites and even games it can be difficult to reproduce the test results as you won't have the same set of test files & scripts (unless Tom makes all this available).
E)Patches to games and other applications can effect performance. WoW now using all cores being a good example. So getting and keeping the exact version of an app becomes critical for a fair comparision. In many cases old releases of commerical applications are not available to the public. Meaning results can not be reproduced by the public once some time as passed. The synthetics are more careful that small patches don't effect performance and generally make old releases available.
E) Some of the above suggestions would not make good CPU benchmarks by virtue of the fact that they are not CPU bound. Microsoft Security Essentials, Office applications and even Visual Studio can often be disk bound rather than CPU bound. Some of the other suggestions like Photoshop, Flash, SETI, etc might also be impacted by the 2D video performance of the video card or networking performance. The synthetic CPU benchmarks don't generally have much if any depenance on the video card & disk.
F) The synthetics can in general use all cores (or a configurable number of cores). By the end of 2011 it will be common place for CPUs to have 8 to 16 cores (with HT). So it is important to have at least some of the benchmarks pushing all core to their limit. What current games are going to use 100% of 16 cores?
G) The synthetics are actually not all that synthetic in many cases (ignoring the simple integer and floating point tests). They often use real world algorithms pulled from real world applications.
So I would be in favour of have some synthetics in the list.
PCMark, SiSoftware, PerformanceTest CPUMark, SuperPi, etc..
For those of us who are software devs, it would be a plus if you guys could also run some database tests and C++/Java compiling performance with either GCC or visual studio. Also, some encryption/decryption tests and to put the icing on the cake, why not the repeat the CPU Stress Test you guys performed 5 years ago...
how about you guys build a free software suite that lets us run your tests on our own computers and uploads the results to the cloud (your cloud) where you accummulate the results, use a majority-vote-style system to eliminate fraud and publish the results for processor performance gathered on YOUR readers' machines?
cheaper for you
exciting for us
everyone wins!
you can still have tests in your own labs but you can better focus those on new/exotic/niche configurations that would make your trusted reliable lab tests 10 times more interesting for us.
again, everyone wins
thanks
cheaper for you
exciting for us
everyone wins!
you can still have tests in your own labs but you can better focus those on new/exotic/niche configurations that would make your trusted reliable lab tests 10 times more interesting for us.
again, everyone wins
thanks
- Status
TRENDING THREADS
-
Question No POST on new AM5 build - - - and the CPU & DRAM lights are on ?
- Started by Uknownflowet
- Replies: 13
-
-
-
-
-
Question Will a 120 watt inverter power my gaming laptop?
- Started by nelska
- Replies: 23
Facebook
Twitter
Instagram