I don't understand. The test is (theoretically) just to see which chip is superior. It's not "cheating" to stack a HT chip against a non-HT AMD chip, since AMD doesn't offer HT.
The problem with the DivX encoding may be that it's being seen as a background process. Since the OS realizes (incorrectly) that it's got four processors, it feels comfortable giving this fourth process free rein of a "processor". Otherwise, it will be starved on purpose. Check the documentation on Win32PrioritySeparation.
EDIT: By the way, even in multi-threaded scenarios HT gives only around a 10% boost at the most. Sometimes, it can actually impede performance on servers performing very specific tasks. It's good for home use almost all the time, though, due to things like virus protection. Those are almost always disabled for these stupid tests, which makes the results very unrealistic.
EDIT: By the way, even in multi-threaded scenarios HT gives only around a 10% boost at the most. Sometimes, it can actually impede performance on servers performing very specific tasks. It's good for home use almost all the time, though, due to things like virus protection. Those are almost always disabled for these stupid tests, which makes the results very unrealistic.
Googer
Lifer
- Nov 11, 2004
- 12,576
- 7
- 81
Because the OS sees it as 4 CPU's and the OS gives each of the 2 CPU's more work to do. It's not an apples to apples compairison. Windows is assigning less work to each of the AMD CPU's.
Also Using your logic, then I would be fare to compaire a quad core chip AMD (if one exisits) to a Dual Core Pentium with HyperThreading. 4 cpu's against 4CPU's, now that is fair.
You people are lame, why is this thread not locked yet?
So the review is biased because they use 4 programs to suit the HT? Well what the hell are they supposed to do? Use only 2 programs so that the potential built into the intel cpus is going unused? Is that a fair benchmark?
Toms simply set these two machines up to run & see what the results are. If anyone can bring forth ANY EVIDENCE to show that their results are unfairly skewed, then do so. If not, then stop spewing crap out of your mouths just because the results don't follow along with your own biases.
So the review is biased because they use 4 programs to suit the HT? Well what the hell are they supposed to do? Use only 2 programs so that the potential built into the intel cpus is going unused? Is that a fair benchmark?
Toms simply set these two machines up to run & see what the results are. If anyone can bring forth ANY EVIDENCE to show that their results are unfairly skewed, then do so. If not, then stop spewing crap out of your mouths just because the results don't follow along with your own biases.
Goober, it doesn't matter. The test is just to see which chip is more capable. It's not "unfair", that's just the way the chips work. If the Cell chip were being compared, it would have many more physical processors than either; it's not "unfair" that IBM and Sony see fit to put so many microprocessors together.
Why the hell would you ever think it was an apples-to-apples comparison?
The two chips have RADICALLY different architectures.NOtice how all these new accounts are starting these blatant troll threads...new user names with a dozen post all in the same thread...Dothan is another touting existing 3ghz dual core dothans at 3 times the speed of current X2's....
I would check to see if they are same person......
The toms test is a model of incompetence, blatant and proven lies (check out cpu/forum), and now cover up....It is conceivable the EE will still be the better multitasker with apps that can take advantage of the 2 virtual cores (HT) for an effective 4 cores.....However in contrast to major review sites most apps dont appear to use the HT well so he may really have had to rig apps (type of apps in conjunction) to get these results...May not be simlar at all with 4 completely different cpu apps....
Also anyone who has done this knows process priority can be effected by order of starting apps, whether an app is in the foreground or minimized to the background....I can skew these results 2-3 different ways....Since the result can be skewed one should really question any results until verified by others......Toms test are already renowned for being impossible to duplicate...
Yes I am calling him incompetent and a Liar....his current test ad the debacle it has been with the Intel p4 being down 1/4 of the time is evidence of that...
I would check to see if they are same person......
The toms test is a model of incompetence, blatant and proven lies (check out cpu/forum), and now cover up....It is conceivable the EE will still be the better multitasker with apps that can take advantage of the 2 virtual cores (HT) for an effective 4 cores.....However in contrast to major review sites most apps dont appear to use the HT well so he may really have had to rig apps (type of apps in conjunction) to get these results...May not be simlar at all with 4 completely different cpu apps....
Also anyone who has done this knows process priority can be effected by order of starting apps, whether an app is in the foreground or minimized to the background....I can skew these results 2-3 different ways....Since the result can be skewed one should really question any results until verified by others......Toms test are already renowned for being impossible to duplicate...
Yes I am calling him incompetent and a Liar....his current test ad the debacle it has been with the Intel p4 being down 1/4 of the time is evidence of that...
Googer
Lifer
- Nov 11, 2004
- 12,576
- 7
- 81
No Bias here, I am typing this message from an Intel Machine and under my roof I have computers with CPU's from IBM, AMD, Intel, and Cyrix. I have no loalty.
The unfareness that the Intel machine has is that the software sees it as 4 CPU's and not two. This would be a better comparison if we had Dual Core opterons on a Dual Socket motherboard, or simply turn off hyperthreading so that THE SOFTWARE ONLY RECOGNIZES 2 CPU'S ON BOTH MACHINES.
This is not a test of equals, Tomshardware has messed up and they should start over.
Personaly I don't care who wins, just as long as the race was won fair and sqare. Because I would buy either AMD or Intel based on these results if they were accurate but they aren't, because Intel Machine has an unfair advantage in the windows operating enviroment
Goober, it is not unfair since the two chips are both aimed at the same market. The comparison may be skewed or unfair (or just suck), but it's not "unfair" to compare two desktop multicore chips. Turning off hyperthreading would not make it more fair; most people (people not just interested in skewing things in AMD's favor) would then complain, since the Intel chip would be hobbled.
The test is supposed to be an unbiased test of two multicore chips running multiple processor-intensive processes at once. You need to think more in terms of "How much will this processor do?", instead of "Why isn't the AMD chip kicking more ass?"
It would not be unfair to stack the Intel chip against four-way Opterons, it would just be a different comparison. However, it would be comparing one multicore chip against four processors, and the two HT virtual processors only give a 10% maximum boost anyway. It'd be a one-dual-core modern chip vs. two dual-core modern chips (the latter on a high-end workstation mobo), and I'd be surprised if the Opterons didn't win. They'd totally suck ass if they didn't. THAT would be apples and oranges.
Don't waste your breath. The AMD zealots here wouldn't be happy if Tom took the EE 840, broke off one core, and somehow managed to cripple it with PC2700 DDR. They'd still scream about something being unfair.
And the Tom's bashing is really pathetic. Agree with Dr. Pabst or not, but don't bash the integrity of the site. These same people were bashing Anand not all that long ago (pre Athlon 64) because of his supposed Intel bias (remember when Intel ads were all over this site?)
It is funny, also, to see AMD zealots crying foul about HT being enabled. The processor supports it, so why not? After all, a lot of you have said the EE 840 actually performs WORSE with HT enabled in many scenarios. So why do you care? You've also stated that AMD "doesn't need" HT. Fine. So why are you whining?
Come on Duvie, do you really assert that Tom has "rigged" the applications? How bad have things gotten?
How in the hell do you "rig" WinRAR and LAME? Or FarCry? Now if you want to speculate that perhaps the thread PRIORITIES are set differently on each machine or something to that effect, fine. But when you say "rig" you are asserting Tom has altered the executable code of these applications to somehow benefit Intel. Do you really expect ANYONE to believe that?
Somehow I suspect that EVEN IF ANAND did the "test" himself, you wouldn't believe it. There'd probably be a few hundred "conspiracy theories" threads floating around about some collaboration between Anand and Thomas Pabst.
Good lord, it is seen as 4 CPUs because of HyperThreading WHICH THE EE 840 SUPPORTS AND THE X2 DOES NOT. So shall we disable SSE2 and SSE3 too? What else can we disable to please you? Give it up. And this test is designed to be REALISTIC. How many people who spend $1000 for a top-end CPU are going to DISABLE one of its most prominent features?
Yeah, let's run the test 100000 times until that X2 finally wins.
Fine, let us benchmark under Linux then. I'm all for it.
Googer
Lifer
- Nov 11, 2004
- 12,576
- 7
- 81
Apples to apples is when the OS sees the same number of CPU's for both systems. Enableing hyperthreading is would be like compairing going back to january and doing and AMD vs Intel SSE3 test. Only Intel could win that one so you would have to disable SSE3 and use only SSE2 for a fair comparison. The same is true with hyperthreading. When ever AMD comes out with their version of it then it will be APPLES TO APPLES.
FACTS:
1.) Each platform should represent the best that platform has to offer. DDR-400 for AMD, DDR2-667 for Intel.
2.) Intel gets more memory bandwidth, AMD ends up with superior timing. Both get what they need the most.
3.) Intel chip comes with hyper-threading, so guess what? It gets hyper-threading.
4.) Intel chip comes with higher clock speed, so guess what? It gets higher clock speed.
5.) AMD has a vastly superior design with less heat and wattage. Guess what it gets.
6.) The video encoding fisaco is not AMD's fault, nor the OS's fault. It's the user's, because the user has the ability to change it.
7.) Due to reliability problemsin addition to the above, this test proves and means nothing.
8.) AMD's dual-core processors still beat Intel's products senseless. No competition at all.
9.) However, I'm sure you could find SOME combination of 4 applications that would make the 840EE work better than the 4800+.
10.) There is no #10, but I thought it a good number to end on.
All fanboys, both sides, shut up. Please.
EDIT: The "Apples to Apples" analogy is horribly flawed. By that logic we should wait until Intel gets rid of netburst, as its contender is crippled by that while AMD's is not.
1.) Each platform should represent the best that platform has to offer. DDR-400 for AMD, DDR2-667 for Intel.
2.) Intel gets more memory bandwidth, AMD ends up with superior timing. Both get what they need the most.
3.) Intel chip comes with hyper-threading, so guess what? It gets hyper-threading.
4.) Intel chip comes with higher clock speed, so guess what? It gets higher clock speed.
5.) AMD has a vastly superior design with less heat and wattage. Guess what it gets.
6.) The video encoding fisaco is not AMD's fault, nor the OS's fault. It's the user's, because the user has the ability to change it.
7.) Due to reliability problemsin addition to the above, this test proves and means nothing.
8.) AMD's dual-core processors still beat Intel's products senseless. No competition at all.
9.) However, I'm sure you could find SOME combination of 4 applications that would make the 840EE work better than the 4800+.
10.) There is no #10, but I thought it a good number to end on.
All fanboys, both sides, shut up. Please.
EDIT: The "Apples to Apples" analogy is horribly flawed. By that logic we should wait until Intel gets rid of netburst, as its contender is crippled by that while AMD's is not.
mechBgon
Super Moderator<br>Elite Member
- Oct 31, 1999
- 30,699
- 1
- 0
Disabling Hyperthreading on the P4 would be silly. You wouldn't buy a 16-cylinder Dodge Viper and then go yank eight of the fuel injectors in order to drag-race an 8-cylinder Corvette at the track. You bought a 16-cylinder engine, you're going to use all of the cylinders all of the time.
'Tard Hardware Guide has certainly messed up their test, though. The reboot discrepancy, the mysterious accumulation of encoding time on a P4 that is at 0% CPU usage... uh, NO. :roll: If AnandTech did the same test, and stuff began crashing and burning, I think AnandTech would pull the test and explain what happened, rather than continuing to use it as a publicity stunt/traffic generator (*cough*).
'Tard Hardware Guide has certainly messed up their test, though. The reboot discrepancy, the mysterious accumulation of encoding time on a P4 that is at 0% CPU usage... uh, NO. :roll: If AnandTech did the same test, and stuff began crashing and burning, I think AnandTech would pull the test and explain what happened, rather than continuing to use it as a publicity stunt/traffic generator (*cough*).
i think that to test it they should run the same amount of apps as there are cores(as the OS recognizes it...), and then when the amd x2 finishes the first two apps starts up the next app, whilst intel pentium ee does all 4 at the same time.... and the amd alternates two apps at a time for however logn thg does their live thang.... and intell does all 4 at once while.... thats my opinion as i am entitled to it....
personally i like both sites at and thg, they both give good reviews and different viewpoints, the AT forums just pwn the shizz outta thg, i cant stand thg forums so ugly and so crappy..... sorry thg you need help there.... and the whole biased thing, well thg does often seem intel fanboy (why not use the word fangirl? has a positive connotation even if they are fans of something you dont like ) the AT forums people liek AMD alot.... and thus fanboys.... w/e i like these forums and you aint gonna stop me from reading em!, banning wouldnt work... i have sources....
personally i like both sites at and thg, they both give good reviews and different viewpoints, the AT forums just pwn the shizz outta thg, i cant stand thg forums so ugly and so crappy..... sorry thg you need help there.... and the whole biased thing, well thg does often seem intel fanboy (why not use the word fangirl? has a positive connotation even if they are fans of something you dont like
) the AT forums people liek AMD alot.... and thus fanboys.... w/e i like these forums and you aint gonna stop me from reading em!, banning wouldnt work... i have sources....
kitkat22
Golden Member
- Feb 10, 2005
- 1,464
- 1,333
- 136
I agree with Chocolate Pi and I'm sorry for continuing the bump, but I had to add my two cents. This test is a test of "one" type of scenario on the two platforms. Upon looking at the current data Intel is thumping AMD in the Divx scores, however take a gander at the other three applications running. It would appear to me that while Intil is hitting AMD in the face AMD is in turn clobbering Intel everywhere else. Basically what I see is unequal distribution to all apps. AMD spends more processes on the first three apps while intel sends it's processes somewhat equal across the map. The issue is scheduling. I would tend to say that if the AMD were to distibute equally it would lie on par with the Intel processor and vice versa. I would say the test is invalid because this is again "one" scenario and can change over different applications and the number of applications. I would have to draw the conclusion that benchmarking any dual-core processor is going to be extremely difficult.
EightySix Four
Diamond Member
- Jul 17, 2004
- 5,122
- 52
- 91
I say they do it on linux so we'll stop having the b!tchin and moanin, can't wine about the OS scheduler there.
Cerb
Elite Member
- Aug 26, 2000
- 17,484
- 33
- 86
6000 RPM fan, when they run cooler than the 130nm single-core CPUs?
...and is that the worst layout for a review in history, or what? I fail to see where it is said that Intels is better. Hell, I don't see where it says anything escept the tests, and then some graphs.
...and is that the worst layout for a review in history, or what? I fail to see where it is said that Intels is better. Hell, I don't see where it says anything escept the tests, and then some graphs.
Originally posted by: Pabster
Good lord, it is seen as 4 CPUs because of HyperThreading WHICH THE EE 840 SUPPORTS AND THE X2 DOES NOT. So shall we disable SSE2 and SSE3 too? What else can we disable to please you? Give it up. And this test is designed to be REALISTIC. How many people who spend $1000 for a top-end CPU are going to DISABLE one of its most prominent features?
Yeah, let's run the test 100000 times until that X2 finally wins.
Fine, let us benchmark under Linux then. I'm all for it.
Man shut the heck up....
< 'Tard Hardware Guide has certainly messed up their test, though. The reboot discrepancy, the mysterious accumulation of encoding time on a P4 that is at 0% CPU usage... uh, NO. If AnandTech did the same test, and stuff began crashing and burning, I think AnandTech would pull the test and explain what happened, rather than continuing to use it as a publicity stunt/traffic generator (*cough*). >
as said by mechbgon and others who have been watching this more thoroughly then you "papsmear"...
If you knew anything you would have also known they screwed up the last stress test as well....I wont go into that story cause I know they are running your beloved preshot....
You dont have to rig the apps you "know nothing"....many who have played with OS and apps know you could manipulate priority, having specific apps in foreground or background effects their priority, and overall how you start an app in certain orders will give precedent over others.....Use your fvcking brain....
You could easily configure the apps so winrar sucked instead of divx...or even lame encoding Cds...lettnig the encoding rock towards Intel I bet was a desired result of Toms....
Also I would never doubt that the EE would not do beter with 4 apps cause I know that EE and the hT can help it...It did in some AT test but also hurt it severely in others...Funny how this one favored it when it could have gone the other way....You want to bet that they tested this out prior to setting up the review??? Look at ATs review if you need some help jarring the cobwebs out of the brain...
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter