THG new review

[Duvie]

Anyone find it odd

http://www.tomshardware.com/stresstest/charts.html

How the AMD averages more fps with 4 apps running last week then it is right now??? It was 35-37 before and I assuem it should be much more now with no apps stealing any cycle from it...it was only using 46% on average in the 4 app test...

What did this moron do now...i would think both fps of both systems should be up....

If you look here

http://www.tomshardware.com/stresstest/load.html

It would seem that this is a video card limitation...doesn't it??? Funny again how it did more before when more apps were running.....I guiess toms is trying to give INtel one of the test since they got HAMMERED by 29% in Winrar, 27.5% in DIVX and 5% in Lame encoding......

This will start the new bashing of Tomshardware....I wont give the germans a pass on this one.....They are the laughing stock and should continue to hold that title...




http://www.techreport.com/reviews/2005q2/athlon64-x2/index.x?pg=5
http://www.xbitlabs.com/articles/cpu/display/pentiumd-820_10.html
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2411&p=2

interesting...high quality and they get 88fps...WTH settings are they running now

http://www.tomshardware.com/cpu/20050509/cual_core_athlon-14.html


EDIT: They never really updated the info after the failed SLI test but they may be still running 1600x1200 with unknown quality settings....I know this has to be a vid card bottleneck.....Not much of a cpu test but more of a vid card and system test...i am scouring other sites since I know Tom"s "Anomaly Testing"....

 

[AnandThenMan]

I've basically given up trying to understand anything about that stress test. More often than not the numbers just don't make sense.

The biggest problem I have with the test is it starting out being a stability test, to see how both platforms held up under extreme load. But Toms decided to "reset" the test several times which really gets on my nerves. How about this THG, give us a nice graph with stats FROM THE BEGINNING. Oh wait you can't because you screwed up so badly none of the numbers make sense, stats are missing, stats change for no apparent reason. And you're right Duvie, the AMD box was getting higher fps with a higher CPU load which seems impossible to explain.

BAH Toms truely is the laughing stock of the community.

 

[blckgrffn]

Amen, and that is all...

 

[Duvie]

I like how they haven't been smart enough to relaunch the Intel's temp reporting for like the last week.....

I think ppl just need a place to see the truth and the truth is they are being rather coy with what the settings are being tested right now and they dont seem to jive to testing done in the first 3 rounds of testing...

 

[Markfw]

Yea, and they said the Intel won the first round, when most of us saw it a a tie ! and now this. The spins continue....

 

[Duvie]

It was virtually a a tie if you just look at fps but one should really be looking at the number of runs....I dont know why they are soooo stupid to compare fps and then run the test for weeks...the best comparison was the number of runs completed....

 

[paadness]

So what was the point in the first place?

 

[Lonyo]

I thought the Intel system had SLI and the AMD system was running a single card?
That could explain some of the difference.

Looks like they are both in single card setup from the pictures.

 

[Duvie]

They both started out with SLI at the very beginning of the stress test but the SLI Intel boards were "being blamed" for the instability...Then for the 2nd and 3rd rounds and these subsequent single test it was no SLI for either...though it was working 100% for the AMD system.....

They are definitely single cards now...Just the funny thing is that it cant even get the same 36-37fps level it was getting when 3 other apps where fighting for cycles...really just 2 since Divx was a low priority....

 

[ZobarStyl]

Why bother to pay any more attention to that test? We've already seen all the results we needed: Windows Scheduling is crappy, and the EE couldn't stay up without massive reworking, which, IIRC, was the same thing that happened last time in Tom's single core stability run ("Oh um the Intel PSU was busted despite being a brand new 400 dollar monster.) They keep doing these tests claiming it's stability they are looking for, but the scientific method apparently means nothing to them. The fact is, had they approached this scientifically, you would never alter multiple variables mid-test, especially if you were testing overall stability. If that's what they truly wanted out of this test, the dual core EE failed with flying colors. If you're testing stability and it breaks, the test is over, and the subject has not passed. If you want to restart the test each time the Intel gets a new motherboard, that's fine. But you cannot rightly say that data from numerous different motherboards is part of a single test.

A analogous situation would be if a car reviewer did roadworthiness tests on two cars, but one's engine broke down after 100 miles, so it got a whole new engine and the test continued: the broken car was not able to go more than 100 miles, and that is the limit found. If the Intel system was only up a handful of days before the motherboard crapped out, then the expected stability is the time up until the system broke.

 

[Duvie]

I do this more so now just to poke fun at the imbeciles...and ofcourse try to keep them honest....

If they ran this at levels I see most review sites test or even run with very HQ with similar vid cards they show a X2 as better or at least show the FX55 kill the 640's so it should not be any much different...Techreport showed in his testing which may be of a different scenes running but it is slightly multithreaded to the tune of about 10% as seen from a 4000+ to a 4800+ X2, and some of that can be the rev E core revisions...So I cant see where the 840 got such a great burst over the 640 which has HT as well...It is clear from test this thing is not accesssing the 4 cores. If anything it is one solid core 100% and a very small part of another...AMD is running like 90% one one and 10% on the other...

 

[HDTVMan]

For the people who dont know any better that go to THG they convince them that Intel has a competing processor. The rumor is he gets a lot of free CPU's from Intel. Note he has or Intel supplies him with AMD cpu's. AMD wont even send him a cpu since he's a sellout.

When one site gets different results from every other site you have to wonder whats wrong. His name is history and it shouldnt be too long before his site dives into nowhere land. His site gets too much publicity here.

 

[RichUK]

THG= Intels lil BIATCH !!!!!!!!!

 

[imported_michaelpatrick33]

How come they didn't leave the SLI mode on AMD if it was proved "stable" in a stress test? Why did they handicap AMD in a "stress" test because of an Intel problem? I don't get it. You would leave the stable system with an advantage because it is "stable". Wasn't that the point? Why did they reset the scores instead of leaving AMD's scores as they were because it was "stable" and Intel wasn't? Shouldn't the downtime count against an unstable platform in a "stability" test? Why were FPS aveaged in the game test if AMD completed many more runs wouldn't that indicate superior performance? How come a "stress" test has a winner based on performance? Wouldn't the winner be the one with the most uptime regardless of performance?

The bias is so obvious that it has become hard for me to even go there to look at any reviews without wondering if they have errored in their approach.

 

[linkgoron]

god...
This is in update 10 in german, i translated it using altavista translator:

Despite the clearly lower clock of the Athlon 64 X2 4800+ with 2.4 GHz opposite the Intel Pentium 840 EE with 3.2 GHz the AMD system can set itself well in scene.

We know that clock speed is what determines speed...

Monday, 20 June 2005: For Friday only in each case one application runs on both systems. Contrary to in former times there is not thus the simultaneous enterprise of several applications no more. We specified the order - the test with the packing of files (WinRAR) began 18 June until Saturday, whereupon the MP3-Encoding of the Michael Jackson CD followed until Sunday, in the connection the DivX Encoding of the James bond DVD was settled and as the latter runs up-to-date the 3D-Spiel Farcry. The live-charts those to be constantly updated, shows a clear picture. A small note at the edge: With the Intel system the virtual multi-processing (HT - Hyperthreading) is naturally activated, so that even four CPUs are available. AMD possesses this feature not and builds on two physical units. Despite the clearly lower clock of the Athlon 64 X2 4800+ with 2.4 GHz opposite the Intel Pentium 840 EE with 3.2 GHz the AMD system can set itself well in scene. The trial balance from three single's applications looks in such a way: Luggages of files with WinRAR: AMD system with 29.5 per cent of increased output MP3-Encoding of the Jackson CD: AMD system with 4.7 per cent of increased output Divx Encoding of the bond DVD: AMD system with 28.2 per cent of increased output until Tuesday 21 June against 18 o'clock remains fourth application, the 3D-Spiel Farcry, actively.

It isn't a great translation, but you can understand what they say if you try...

 

[imported_michaelpatrick33]

Originally posted by: linkgoron
god...
This is in update 10 in german, i translated it using altavista translator:

Despite the clearly lower clock of the Athlon 64 X2 4800+ with 2.4 GHz opposite the Intel Pentium 840 EE with 3.2 GHz the AMD system can set itself well in scene.

We know that clock speed is what determines speed...

Monday, 20 June 2005: For Friday only in each case one application runs on both systems. Contrary to in former times there is not thus the simultaneous enterprise of several applications no more. We specified the order - the test with the packing of files (WinRAR) began 18 June until Saturday, whereupon the MP3-Encoding of the Michael Jackson CD followed until Sunday, in the connection the DivX Encoding of the James bond DVD was settled and as the latter runs up-to-date the 3D-Spiel Farcry. The live-charts those to be constantly updated, shows a clear picture. A small note at the edge: With the Intel system the virtual multi-processing (HT - Hyperthreading) is naturally activated, so that even four CPUs are available. AMD possesses this feature not and builds on two physical units. Despite the clearly lower clock of the Athlon 64 X2 4800+ with 2.4 GHz opposite the Intel Pentium 840 EE with 3.2 GHz the AMD system can set itself well in scene. The trial balance from three single's applications looks in such a way: Luggages of files with WinRAR: AMD system with 29.5 per cent of increased output MP3-Encoding of the Jackson CD: AMD system with 4.7 per cent of increased output Divx Encoding of the bond DVD: AMD system with 28.2 per cent of increased output until Tuesday 21 June against 18 o'clock remains fourth application, the 3D-Spiel Farcry, actively.

It isn't a great translation, but you can understand what they say if you try...

Did a technical review site really say Despite the clearly lower clock of the Athlon 64 X2 4800+ with 2.4 GHz opposite the Intel Pentium 840 EE with 3.2 GHz the AMD system can set itself well in scene.

Well, that one sentence exposes Tomshardware as a pro Intel site that is attempting to cater the neophytes toward the Intel "higher clockspeed". That is a reprehensible way to phrase the AMD solution in a technical "stress" test geared toward the computer savvy.

Clearly lower clockspeed? Clearly lower clockspeed? Clearly lower clockspeed? Clearly lower clockspeed? Sitting in corner giggling and chewing on carpet in stupefaction at such a statement. Clearly lower clockspeed? Clearly lower clockspeed? LOL

Ridiculous, shocking, hilarious and ridiculous
I can sum up that line and what it represents to me only with the this profound picture :roll: :shocked: :laugh: :roll:

 

[sandorski]

Bleh, I'm not even going to go look at the fiasco in progress. He's probably getting more hits and making more Cash now than he would for months worth of less controversial Reviews. Don't make him(website) Rich(er).

 

[Keysplayr]

I think that the only thing that can be said that nobody can even begin to doubt is, nobody truly knows what they "are" or "are not" up to. For all we know, they are incompetant, yes. If you enter this discussion with the idea in your head that Toms is Intel biased, then you should think just the opposite becase they undoubtedly have shown that AMD is more mature, stable, faster than Intel.

If Tom's hardware was truly anti AMD and pro intel, I am sure they would not have shown the Intel platform in such a disastrous way. Don't you think?

What I do think is that all of you are kind of blowing this whole think WAAAAYYYY out of proportion. Everyone is like "What the Hell is Tom's trying to pull now" or "You guys notice this?" or "Funny how they took a sharpened flathead screwdriver and drove it through the heart of the X2 chip with a 6 pound lump hammer to slow it down and make intel look better." It's lame. stop it.

 

[Viditor]

Originally posted by: keysplayr2003
If Tom's hardware was truly anti AMD and pro intel, I am sure they would not have shown the Intel platform in such a disastrous way. Don't you think?

Unless they were incompetent as well...

Actually, I've often thought that it was no coincidence that THG moved their headquarters to within a few minutes walk from Intel's HQ...

 

[Sentential]

Im going to try and be as objective as possible with this.

First off there is no question that the nforce4 for intel is absolutly horrible. It has to be one of THE worst motherboards I have ever seen or heard of. That being said Im sure that SLI actually DID cause a conflict (thats how bad it is).

Second Intel's dual cores are a real nightmare if you arent prepared for them. Cooling and the FETs take a serious beating. However IF you get all of that under control, they do just fine. HOWEVER I would NEVER run it on stock cooling. Nothing less than water-cooling or an XP120 should be used, period.

Also besides some flukes, in terms of stability Intel is marginally ahead. Even dispite how badly THG botched it and made it appear. On that note I gotta be honest the more I play with both systems, the more they appear to be about the same.

However, on CPU intensive stuff there is no doubt that the AMD system should/will pull ahead. There is nothing that Intel will be able to offer (from what I see) that can honestally beat AMD until mid 2007 with the release of Conroe. Until then they will be playing catchup.

In addition I think THG's handeling of this benchmark is just aweful. Skewed results (that duvie pointed out) and other garbage. Frankly they should be ashamed and should give up while they can.

 

[Stoneburner]

It's nice that people on these forums can point out flaws so easily, but shouldn't some official website (like anandtech maybe?) either try to run a similar test or even, god forbid, call out THG? I'd hate to see THG be allowed to pull a fox news.

 

[Pabster]

Originally posted by: Stoneburner
It's nice that people on these forums can point out flaws so easily, but shouldn't some official website (like anandtech maybe?) either try to run a similar test or even, god forbid, call out THG? I'd hate to see THG be allowed to pull a fox news.

You can bash THG all you want but don't start about FOX!

 

[Markfw]

Originally posted by: Sentential
Im going to try and be as objective as possible with this.

First off there is no question that the nforce4 for intel is absolutly horrible. It has to be one of THE worst motherboards I have ever seen or heard of. That being said Im sure that SLI actually DID cause a conflict (thats how bad it is).

Second Intel's dual cores are a real nightmare if you arent prepared for them. Cooling and the FETs take a serious beating. However IF you get all of that under control, they do just fine. HOWEVER I would NEVER run it on stock cooling. Nothing less than water-cooling or an XP120 should be used, period.

Also besides some flukes, in terms of stability Intel is marginally ahead. Even dispite how badly THG botched it and made it appear. On that note I gotta be honest the more I play with both systems, the more they appear to be about the same.

However, on CPU intensive stuff there is no doubt that the AMD system should/will pull ahead. There is nothing that Intel will be able to offer (from what I see) that can honestally beat AMD until mid 2007 with the release of Conroe. Until then they will be playing catchup.

In addition I think THG's handeling of this benchmark is just aweful. Skewed results (that duvie pointed out) and other garbage. Frankly they should be ashamed and should give up while they can.

I agree with most of what you said, but with the AMD system with ZERO re-boots, how can you possibly say Intel is ahead in stability ??????

 

[Viditor]

Originally posted by: Sentential
Also besides some flukes, in terms of stability Intel is marginally ahead. Even dispite how badly THG botched it and made it appear. On that note I gotta be honest the more I play with both systems, the more they appear to be about the same.

Ahhh stability...that old chestnut!
Let's look at the whats and wherefores of system stability...

1. Firstly, we need to remember that we are talking about system stability. If you looked at just the CPU, you would have to give AMD the nod on stability based on heat alone (the higher the heat, the lower the stability).
2. Second, many folks are of the opinion that Intel is more stable because they manufacture their own mobos, chipsets, etc... while AMD must rely on 3rd parties.
The fact is that both of these aren't quite true... Intel has a large number of the Intel branded mobos manufactured by Foxcomm (just as Asus has had some of their cheaper mobos made by PC Chips). Despite what it says on the label, it's near impossible to know who really made the mobo unless you're on the inside.
As to chipsets, I'm sure it will surprise many that AMD is actually one of the largest manufacturers of AMD chipsets... If you look at server boards for example, nearly ALL of them are at least partially AMD chipsets.
Just as Intel has engineers checking the Foxcomm made mobos, AMD has engineers in place at the 3rd party chipset manufacturers.
3. Intel has a reputation for stability that they have earned over many decades. Whether it is as true today as it was then is irrelevant...the perception of stability is probably just as critical as the actual stability itself!

All in all from a technical viewpoint, I would say that AMD probably has the more stable platform...if for no other reason than the far lower heat levels inside.

 

[Markfw]

Well said Viditor !