Socket 939 Sempron found........

[Duvie]

HT still did better cause it at this same point had a bigger lead in both CD encoding and Divx.....It is performing exactly what the EE should (with the (2) virtual cores adding about 20-25% as seen in past HT gains with the prescott cores), OFF....


I dont hink it is performing marvelously yet....

It loses big to AMD in Winrar...almost identical to what it was against the EE

It stays closer in gaming...This test is a bit short so far but the INtel suffered from large dips in gaming going from 35-36fps to 27-28fps in usually 6 hour burst in the test with ht on...

CD encoding is doing worse about half as much as the EE one did...It usually gained <1 per hour and here we are 7 hours in and it has a lead of 2.....

Divx is the big loser as it is clear the low priority is acting just like the AMD cpu and waiting for a handout of spare cycles....I bet both will perform rather close...


Bottom line

Intel wins CD encoding by 7-10%
AMD wins winrar by 30-35%
AMD wins gaming by 5-7%
Intel/AMD tie (so far) in Divx encoding....

 

[Duvie]

If you look at the load balance chart and can find it when it actually registers all apps (both intel and AMD frequently do not eventhough cpu usage is pegged at 100%) you will see that on average AMD applies less percenatge for FarCry and lame encoder...

In FarCry (gaming) and AMD advantage area you can see 4-6% less usage and it still wins...INtel is strong still at encoding and so AMD giving less cycles to Lame gives INtel the advanatge...IN Winrar Intel gives it the less and performance is obvious...

In Divx both over the last 3-4 hours have been real close with both flopping back an forth taking the lead in most % given...Usually this has never topped 1% but tends to stay in the .2-.6% range.....Barn burners!!!

 

[Duvie]

ONce again the network is down on the AMD machine...so once again no numbers have been updated for about 30-40min....skews the data for a time being...system seems to be running fine according to the cpu usage on the graphs....

Again I have never seen someone have so much damn problems with their network....

 

[CrystalBay]

Originally posted by: GuitarDaddy
Although this online test has been a miserable failure up to this point, I think it has opened a lot of peoples eyes about the methods or lack of that THG uses during their review process. And the total lack of credibility in there motives and conclusions. I understand better now why some of Toms reviews seem so far off the mark and don't agree with other credible reviewers results. They have absolutely no clue about control procedures, and their tests are obviously designed with a particular result in mind. If they don't get the result they expected, they just change the test and start over:Q

Lol, particularly when THG alphabetically puts Intel in front of @MD in all test summaries.

has anyone else noticed it ?

 

[boran]

Thats kind of stretching it a bit imho.

And it looks like I was right, the DivX thread is set to a lower priority, so the windows scheduler is flawed when 2 HT cores are active.

 

[Duvie]

Has anyone noticed the difference on the board?? Isnt supposed to be the same Asus mobo the last test ended with?? You would think but if it is it now has a huge modded Northbridge HSF on it...It is huge!!!!

So mod one system cause you are afraid it will crash....I am sure he would not have gone to those lengths for the AMD system considering it id the only one he is showing with a cpu temp and yet all along he has been using a HSF that is not even the standard one that will be boxed with the retail chips....




Also I ma gonna be pissed 12 hours later if the Divx truns to 20minutes now while the AMD system is not reporting data due to network issue so I possibly wont be able to tell who got there first...It could take another 12 hours....I am trying to analyze data and these dumb-asses keep screwing me up....

 

[Duvie]

nearly 4 hours later and those stupid fvcks cant fix the network stream on the Amd chip...i mean cmon you stupid arse germans...It is like 11:45am your dumb-asses have to be in the "lab".....This has got to be a bunch of junior high kids working out of a converted garage or something.....

Some INtel fanboy noob probably will be seeing the scores anytime now and come in raving how the 840 is winning all the benches now...LOL!!!! Probably Tom's goal....



HEY KEYS!!!!

Change the topic from "live streaming data" to 24hour streaming data for INtel cpu and 20 hour streaming data for the AMD.....It was down like 3-4 hours yesterday as well....I dont think the Intel has ever had this network issue with it...just that it didn't register a network connection for like nearly the entire 2nd round of testing.

 

[Jeff7181]

Has anyone come up with an explanation as to why the Intel system is still measuring performance on 4 logical processors?

 

[Markfw]

They did something.... Now the AMD is creaming the intel at Dix, and winning 3/4 easily, but loosing at gaming !!!!! (frames holding though)

 

[Jeff7181]

Originally posted by: Markfw900
They did something.... Now the AMD is creaming the intel at Dix, and winning 3/4 easily, but loosing at gaming !!!!! (frames holding though)

Since so far the MP3 encoding has been pretty even, and now AMD is pulling ahead by a good margin, I think they probably changed task priority.

 

[Duvie]

I think the gaming was off (literally) for awhile leaving more cycles to other things....it would explain more cycles, but I believ the way priorities would work and we still have just 2 cores Duvx woudl still move at a crawl and Lame and Winrar woudl share the extra load gaming was taking...That makes sense on the huge leads it built there. Though winrar now is at same precentages so it may have not got any extra cycles.

I mean look at how long it took those idiots to fix the streaming data thing.....

They fvcked up again and screwed the data yet again.....


I would just give up if I was them



Again someone comment how they put that huge passive (like a Zalman) coller on the NB chipset...So we are doing mods now to just one of the boards???


Also notice how Intel cardiac rhythm is gone??? AMD has a few now and then...It had several in a row but I noticed that only seems to happen in conjunction with the streaming data going down...It hapens so frequnetly with this guys I have been able to see it a few times to tell...

 

[Duvie]

Also notice how AMD's high temp has moved up to 56-57c for max something it didn't manage to do for over 240+ hours of stability testing....I wonder what they are doing there...Lab changing temps?? Too hard to tell cause since they wont enable the iNtel temp we cant tell if they are experiencing smae increase in temps...

 

[Duvie]

Whats weird and I have been looking at the numbers for just the last 30 inutes is tht Farcry is up about 6-10% in cpu usage as wel as winrar and lame is taking it in the shorts....Very different loads based on what was showing the first 12 hours I watched...I wonder why the INtel in all of these test seems to suffer from such randomness in usage patterns...

 

[Pabster]

Originally posted by: Duvie
Whats weird and I have been looking at the numbers for just the last 30 inutes is tht Farcry is up about 6-10% in cpu usage as wel as winrar and lame is taking it in the shorts....Very different loads based on what was showing the first 12 hours I watched...I wonder why the INtel in all of these test seems to suffer from such randomness in usage patterns...

Kinda puts a dent in the "Tom rigged the test for Intel!" theory, eh?

 

[andrewln]

they should really make a graph for the load balancing, so we can see what changed for the past few hour

 

[Duvie]

Well I can see your point there but if you see how they summed up their analysis you can see they draw odd conclusions (spin), and likely would no matter the results....Like saying go with iNtel if you need to do encoding cause they cant seem to see the obvious, they would say hey look AMD's strong point gaming loses out in runs....How are they the only ones so dumb they cant figure out it was the low priority and exactly what I said weeks ago....I mean I am quite embarassed at the lack of knowledge over there....

 

[SlitheryDee]

Originally posted by: Duvie
Has anyone noticed the difference on the board?? Isnt supposed to be the same Asus mobo the last test ended with?? You would think but if it is it now has a huge modded Northbridge HSF on it...It is huge!!!!

So mod one system cause you are afraid it will crash....I am sure he would not have gone to those lengths for the AMD system considering it id the only one he is showing with a cpu temp and yet all along he has been using a HSF that is not even the standard one that will be boxed with the retail chips....




Also I ma gonna be pissed 12 hours later if the Divx truns to 20minutes now while the AMD system is not reporting data due to network issue so I possibly wont be able to tell who got there first...It could take another 12 hours....I am trying to analyze data and these dumb-asses keep screwing me up....

Someone mentioned a while back that HT couldn't be disabled on the asus board. I'll have to look this up...

 

[TGS]

Originally posted by: Duvie

Dont be a moron!!


Obviously the chart has something not working cause lame encoder is obviously getting work done equivalent to Intel right now...

I dont care what the percentage load says but that the cpu (1&2) stays loaded at 100% (which there are you genius) and what the totals are....

Alright, I'm glad everyone caught the fact I was pointing out the inconsistancies on the chart. As the numbers clearly show over 100% overall utilization against both machines from time to time.

Oh wait the responses didn't catch that and just went around tossing about wanton statements.

Next time I'll draw up a mspaint picture to illustrate, visually, what I'm trying to point out. Perhaps a thermometer with the mecury spraying out of the top will correctly show the point I was trying to make apparent.

Of course the "boo hiss" didn't add any satirical element to the post to which I *really* thought someone might find humurous, and thus further review the fact that overall cpu usage was over the technical limit of 100%.

 

[Duvie]

In the last hour...

AMD has increased its CD encoding lead to 19 from 18....

AMD has increased its winrar lead form 180 to 189

AMD has decreased INtel lead in runs from 47 to 45

AMD held 40min lead on encoding....

Looking good for AMD...

 

[Zebo]

Yea and the networks down so howz it tableating....Toms is a fraud.

 

[boran]

From toms 7th update (emphasis added by me):

As announced previously, we will examine the unequal load balancing of the AMD system within DivX encoding activities. We have restarted both systems and shut down Hyperthreading in the Intel system - resulting in only two physical CPU units, matching the AMD system. Perhaps we will get an answer why the AMD system allocates a low priority to Divx encoding. Perhaps this effect is caused by the operating system and re-scheduling events as a result of different core and application numbers.

I'd think they would have seen the light right now it also happens with the intel proc. in fact atm the intel has done waaay less divX encoding than the AMD processor. Not that it is a blame on the intel proc, it is expected and supposed to behave like this, it is a remark to THG to actually understand the numbers they produce. Because at this moment they really dont understand anything of the numbers they produce

 

[imported_Frallan]

This has to be a joke???

Is the test still alive??? I just read the 6th and 7th update and I just cant believe that toms persist. It is half way nteresting but does it make sense any more. I just looked in on the chart and AMD is ahead on everything BUT the game demo ????

I just dont get it...

 

[biostud]

It seems that windows isn't very good at distributing CPU power, when it's running more intense CPU than cores present (virtual or physical). And that intel has found away around this by implementing Hyper Threading.

 

[TGS]

So to sum up and let the thread die.

HT > Windows Scheduler


:beer: FIN :beer:

 

[boran]

well no, that is a wrong conclusion. the conclusion is that HT in a way lets the OS bypass thread priorities, something you dont want. because an app you have set to idle will steal away CPU time from a thread that you want to have preference over it. (a game for instance)