Tom's review??? What about anandtech's???

[Duvie]

I may take some heat for this, but I am tired of the ppl here attacking one review site when right under our noises I don't think any great cpu reviews are being done here...


I mean cmon...

Anandtech's Athlon64 3000+ and 3400+ reviews are IMO horrible...why don't we call this a gaming review site and throw out hardware....The test are abyssmal....2 business apps, 1 encoding app, 2 rendering charts but one app....All the rest is pages upon pages of gaming test and synthetic results....

Now tom's site you bash but who else has ran a test that has lots of gaming test but then also scientific, rendering, workstation, encoding of video and audio, etc???

In this case tom's was by far a more thorough review that at least a person can draw some conclusions from...one cannot from anandtech's other then the athlon64 rocks in price and gaming....


I expect this from gaming review sites, but not from here...the cpu reviews are really starting to lack...

 

[lifeguard1999]

Personally I wish there had been more talk about "Cool & Quiet". THG mentions that no one supports it, yet Ace's seems to be able to get it to work.

 

[THUGSROOK]

again ~ at the mercy of irrellivant benchmarks, especially when comparing AMD to Intel.

remember the days (p3) when 3dmark2000 could tell you all you needed to know?
...those days are long gone.


i thought the H|OCP review was pretty good.

 

[zShowtimez]

The problem I see is just the fact that processors arnt the same as each other. Apples to oranges and such. Its getting to a point that we cant really compare them, so many different features on each. Its hard to think about it and come up unbiased. I dunno, maybe adding piles of benches from many different things will work for now.

 

[DivideBYZero]

Have Aces reviewed it yet? They nearly always put em thru the ringer.

 

[DivideBYZero]

OMG I just checked Aces. Their 'preview' is shite.

 

[myocardia]

Duvie, you seem to have forgotten that you are the minority here. Nearly everyone that's interested in OCing, are interested because they are gamers. The fact that you don't game with your pc is fine, but when you're reading a review on a site that's predominantly visited by gamers, how could you possibly be surprised that they use gaming benchmarks?

edit: By the way, those benchmarks that you performed on HTing have almost convinced me that I need to switch back to the dark side! I'm glad they stickied that forum, because it is a very good read.

 

[BigJ]

As mentioned before Duvie, check out Tom's memory timings for the P4 vs. the A64. 2-2-2-5 vs. all of these diff timings for the A64 boards:
2.5-3-3-6
2.5-3-3-7
2-4-4-8

Hell, way to screw with the results. And also as myocardia said, most overclockers tend to be gamers. You really don't want to overclock a system in an environment where an unstable CPU could crash and corrupt a project that was being made for $1000. It's really not the end of the world if a game locks up or crashes. Even a "completely stable" overclock may result in a computer crashing.

Also, IMHO, Anandtech has been slipping lately with articles and reviews in general. Sure that video card roundup they had was very good, but the number of reviews and the quality of them has not been what I've expected from Anandtech.

 

[FPSguy]

I agree with Duvie. Plus, I think Tom's review provides better insight into the gaming strengths and weaknesses of the chips as well.

I for one am glad to have switched back from AMD to Intel for some of the reasons shown in Tom's reviews. Basically, the P4 is a very fast chip -- plenty fast for my gaming -- and it's hyperthreading rocks in certain situations (including, in my opinion, everyday use when more than one application is running). You get the sense of this from using the chips, and from Tom's review. Less so from AT's review.

As Dennis Miller would say, that's just my opinion -- I could be wrong ...

 

[BDSM]

It's very hard to put together a set of benchmarks that will be considered to be "fair". Since the P4 and the A64 are such completely different architectures they will perform very differently in different apps.

I think it's pretty much pointless discussing why this site used those benches and that used those other benches...

What is worth mentioning is when sites intentionally cripple one system. Like THG did with the ram timings. That is in my opinion CHEATING and proves his bias towards Intel without any ANY doubt.

Oh..I'm not an amd fanboy.. my next system may very well have Intel inside.

 

[Duvie]

You are right BDSM that is why I look at many and take into account what they did to the system....

However this isn't about why tom's uses this divx benchmark versus that one....It is about anandtech not running enough types of benchmarks versus the overwhelmin syntheic and gaming benches ans then declare amd64's as best cpu.....They need to clarify versus their incomplete review they need to say gaming champs cause they did not test enough to say anything else IMO....


Lets just try not to make over half of them be gaming....Lets look at gaming as one subtype....multimedia/encoding another ubtype...rendering/workstation another subtype...business apps another....see my point gaming result total should not make up more then 1/2 of all test.....It is just pure lame!!!

 

[BDSM]

Well.. I have six close friends who all use computers. I have asked them what they use their puters for.
We all do gaming and general use ofcourse. I never encode mp3 nowadays cuz everything is already available online long before I get to buy the cd and the same thing goes for my friends... Four years ago all I did was encoding mp3.. Now I do less than..... well.... I just don't encode anymore!

The same thing goes for divx. I'm sure there are some die hard encoders out there that do nothing but encode but they are very few. Only one of my friends encode videos (k-vcd) on a regular basis.

So.. personally I think gaming should definitely weigh very heavy among the benchmarks as it is not only the most demanding task but also a very common one.

I don't know a soul that does 3d rendering personally and I bet that less than one out of a hundred does.. and even less ppl use autocad and pro engieneer.. So those benches are pretty much pointless.


What I would like to see is more of LOWEST framerate in games.. Whatever happened to measuring that?
I would also like to see more multitasking tests as general usage is what we do most of the time.. and everyone loves a snappy system!

So... personally I think Anand is right to incorporate a lot of gaming benches.

 

[Elcs]

Originally posted by: BDSM
It's very hard to put together a set of benchmarks that will be considered to be "fair". Since the P4 and the A64 are such completely different architectures they will perform very differently in different apps.

Not putting anyones views down here but there will always be people complaining about how hardware review sites perform their benchmarks. One of the major problems is the huge differences in the processors themselves. Programmers have a number of instruction sets from either side which they can choose to code into programs. A Processor is bound to perform better with programs that take advantage or better advantage of their capabilities.

Its impossible to perfectly balance bias but more effort could be made by the forerunners and most respected sites around.

 

[OddTSi]

Originally posted by: BigJ
As mentioned before Duvie, check out Tom's memory timings for the P4 vs. the A64. 2-2-2-5 vs. all of these diff timings for the A64 boards:
2.5-3-3-6
2.5-3-3-7
2-4-4-8

Hell, way to screw with the results.

How much of a difference would those memory timings make? If I remember correctly (it was a while ago, so maybe my memory is fuzzy) THUGSROOK did a thread testing a BUNCH of different memory timings and overall the performance increase by lowering the memory timings was nothing to write home about. It seems like people are making a huge deal out of lower memory timings for little to no reason at all. Again, it's been a while since that thread was posted so I might be remembering this wrong, if someone could find it (I've searched to no avail) can you please post a link to it.

And as far as reviews on AnandTech go, I completely agree with Duvie here. The CPU benchmarks are looking more and more like the video card benchmarks (i.e. games, games, games, and wait, more games). And I also agree with BigJ when he said that AnandTech has lately dropped the ball on ALL reviews, not just CPU reviews.

 

[Duvie]

Originally posted by: BDSM
Well.. I have six close friends who all use computers. I have asked them what they use their puters for.
We all do gaming and general use ofcourse. I never encode mp3 nowadays cuz everything is already available online long before I get to buy the cd and the same thing goes for my friends... Four years ago all I did was encoding mp3.. Now I do less than..... well.... I just don't encode anymore!

The same thing goes for divx. I'm sure there are some die hard encoders out there that do nothing but encode but they are very few. Only one of my friends encode videos (k-vcd) on a regular basis.

So.. personally I think gaming should definitely weigh very heavy among the benchmarks as it is not only the most demanding task but also a very common one.

I don't know a soul that does 3d rendering personally and I bet that less than one out of a hundred does.. and even less ppl use autocad and pro engieneer.. So those benches are pretty much pointless.


What I would like to see is more of LOWEST framerate in games.. Whatever happened to measuring that?
I would also like to see more multitasking tests as general usage is what we do most of the time.. and everyone loves a snappy system!


So... personally I think Anand is right to incorporate a lot of gaming benches.

Hardly scientific....However I have 5 relatives I have built new pcs this year and none of them game or game solely....I also have a guy I just built a system for and he games a bit but when he sees me doing the multimedia stuff it inspires him to use that DV camcorder and make dvd quality movies with created menus, etc...

I also don't know what you are talking about many still use divx and many more are actually getting online using other mpeg4 standards as many are doing their own home movies and sharing them with friends and relatives....


I agree you have the right to feel personally that anand can use a preponderance of gaming apps, but again I thought this was a hardware site where we talk about other things then gaming...maybe I ws wrong maybe I should just avoid anandtech's reviews as for GAMERS ONLY....Cause they lack otherwise....

 

[myocardia]

Originally posted by: OddTSi

Originally posted by: BigJ
As mentioned before Duvie, check out Tom's memory timings for the P4 vs. the A64. 2-2-2-5 vs. all of these diff timings for the A64 boards:
2.5-3-3-6
2.5-3-3-7
2-4-4-8

Hell, way to screw with the results.

How much of a difference would those memory timings make? If I remember correctly (it was a while ago, so maybe my memory is fuzzy) THUGSROOK did a thread testing a BUNCH of different memory timings and overall the performance increase by lowering the memory timings was nothing to write home about. It seems like people are making a huge deal out of lower memory timings for little to no reason at all. Again, it's been a while since that thread was posted so I might be remembering this wrong, if someone could find it (I've searched to no avail) can you please post a link to it.

And as far as reviews on AnandTech go, I completely agree with Duvie here. The CPU benchmarks are looking more and more like the video card benchmarks (i.e. games, games, games, and wait, more games). And I also agree with BigJ when he said that AnandTech has lately dropped the ball on ALL reviews, not just CPU reviews.

I don't remember any post like that. I do remember the one to prove that low latencies, at 5:4 or even 3:2, was better than high latencies at 1:1, even for the P4's. I think Duvie did that one, but maybe not. By the way, what would you expect them to use to test how fast a video card is? The only things that make a difference in the speed of a video card are games and CAD, and what percentage of the computer-using population do you figure uses their comp. for CAD??

 

[THUGSROOK]

Originally posted by: OddTSi
...THUGSROOK did a thread testing a BUNCH of different memory timings and overall the performance increase by lowering the memory timings was nothing to write home about.

that was awhile ago.
even my more recent testing on dual channel isnt worth talking about due to the benchmarks i used. (same old story there)

again, the problem is "what benchmarks to use"?
the 2nd problem is that i run a 2.4b system that basically has nothing in common with a 2.4c system when it comes to tweaking. so my results are only relivant the B chip OCers.

 

[BDSM]

Duvie.. My friends were just a lil example. At work I am for some reason considered by most ppl to be some kind of an expert on computers just because I know a lil bit more than your average joe. I therefore regularly get ppl I don't even know of ask me questions on various problems and let me tell you.. they are usually something like: Why won't this game run?".. or why does my computer do this and that in a particular situation... it's the usual stuff. I have never had one single question about video encoding, EVER.

If you go to your local tv/stereo/computer dealer and listen to the jargon of the questions ppl ask the salesmen and the salesmens answers you will realize that pretty much noone wants to encode video with them. BUT.. they all want to be able to play games.. games, games!

You cant' seriously think that most ppl are encoding personally recorded material back and forth different formats on their PC's do ya?.. C'mon.. most ppl have no clue what so ever about stuff like that.

You mention that 5 of your relatives have new puters and that noone of them uses them solely for games. I am sure they do general usage as well. You don't mention anything about video encoding so I take it they don't use it for that, then.

Oh yeah.. I didn't say ppl don't use divx nowadays.. Sure they do.. but they watch movies they downloaded.. not so much encoding..

Ok.. sorry for my fragmented posts.. But.. what I am trying to say is basically that very few of the general population encode divx videos, use autocad or 3d studio on their puters while most of them do gaming!


Regarding timings.. Well. I don't know exactly how much of a difference latency differences make in real life apps. Though in sandra mem bench for me 2-2-2-5 yields 3,05% (int) and 3,95% (fpu) better results than 2,5-4-4-8.

That may not sound very dramatic but 3-4 % is about one speed grade for processors in many games.. Not that the latency diff would make that big a difference.

But the big issue on tom using "tweaked" timings on the amd rigs here is that when ya compare two products you will always want to isolate their performance by testing them under as similar conditions as possible and he didn't!

If ya start tweaking the results you will lose trust veeeery fast.. ... After all.. who says he didn't do ALOT more tweaking than just the ram latencies!?

ok.. I hope u all understand what I am trying to say now.

 

[Duvie]

The 5 relatives do a multitude of different uses...

1 does mainly photo editing....

1 does do video archiving of old vhs tapes and uses their camcorder to download and creat DISK THAT PLAY IN dvdS...

2 of them general uses, kids may play a game here and there but nothing extreme graphics couldn't handle...

1 the last one is new to the whole game but want to do it all...just bought a spendy dig camcorder and dvd burner for it...his son will use it as he is going to go to graphic artist school here shortly....

I am just saying that is a lot of non gamers...Don't forget about us or definitely pick your wording of your conclusion better to not declare one cpu over the other as superior when all you tested was gaming....


I don't want to argue over this topic.....This site for me is just not the site I will follow for getting my cpu reviews...unless I start gaming....I wouldn't count on it though...That is what a PS2 is for IMO...

Hey I do play CSI on my system!!!

 

[BDSM]

We definitely need a poll. But it would be pointless here because anandtech readers hardly represent the general population.

 

[Jeff7181]

Duvie... exactly how is this inaccurate...

AMD is still the best bet when it comes to business/gaming/2D workstation either from a performance perspective or because of a superior price/performance ratio, while Intel offers the best in encoding and 3D rendering performance.

???

*EDIT*
I don't think many people here need to be told this. This is the way it's been for years... Intel for multi-media, AMD for gaming/business. That's the nature of the beasts. AMD's architecture is more suited for gaming and office type workstations roles... and Intel's architecture is more suited for multi-media and rendering workstations. And neither one performs that badly in the other's strengths that make it unusable. If the computer you're putting together will be used mainly for gaming, right now the smart choice is an AMD 64 system... if it's mainly used for multi-media type applications, the smart choice is an Intel system.

The thing is... most people either do a lot of video/audio encoding, or they do VERY little. Which IMHO makes an AMD 64 system a better fit for their needs.

 

[DAPUNISHER]

I agree with one of BDSM's points, and that's about Tom's deserving criticism for their lack of reputable testing methodolgy. Just a quick example without searching all the threads that have been posted in these forums that clearly demonstrate questionable practices would be the following. In his A64 chipset shootout a couple days ago he used older drivers with the SiS755 chipset than Anandtech did in their review back in November! If you have a look here from the AT review link it states

The Gaming performance of the SiS755 Reference Board equals or tops the best that we have seen so far on an Athlon64 chipset. SiS has been hampered in the past with AGP driver issues, which have impacted their scores on video-intensive benchmarks. Gaming scores certainly indicate that the new 755 chipset combined with 1.17a AGP drivers are a huge improvement and comparable to the best video performance we have seen.

Now if you check Tom's test system config here you'll see he used older 1.16A AGP&2.03 IDE despite having just posted the review days ago, why? It obviously impacted the performance of the SiS reference board as it wins every other A64 shootout except the OCWorkbench in which the SOYO Dragon Ultra with AMD8151 matched up nicely overall, and in all fairness the SiS755 hasn't shown up from an Enthusiast's board maker like Abit or Asus yet , which typically outperforms the reference board and offers even more overclockability. Anywho, my point is that while the lack of significant HT testing does makes all the latest P4C vs A64 reviews *not just Tom's and Anand's* conclusions rather broad sweeping and potentially erroneous, Tom's lack of consistancy and dubious testing methodology is well deserving of all the flaming he recieves IMHO.

 

[Duvie]

Originally posted by: Jeff7181
Duvie... exactly how is this inaccurate...

AMD is still the best bet when it comes to business/gaming/2D workstation either from a performance perspective or because of a superior price/performance ratio, while Intel offers the best in encoding and 3D rendering performance.

???

*EDIT*
I don't think many people here need to be told this. This is the way it's been for years... Intel for multi-media, AMD for gaming/business. That's the nature of the beasts. AMD's architecture is more suited for gaming and office type workstations roles... and Intel's architecture is more suited for multi-media and rendering workstations. And neither one performs that badly in the other's strengths that make it unusable. If the computer you're putting together will be used mainly for gaming, right now the smart choice is an AMD 64 system... if it's mainly used for multi-media type applications, the smart choice is an Intel system.

The thing is... most people either do a lot of video/audio encoding, or they do VERY little. Which IMHO makes an AMD 64 system a better fit for their needs.

Who said I was saying anandtech's conclusions were inaccurate...i said they have a lopsided suck-arse review....That is what I said..

 

[Jeff7181]

Originally posted by: Duvie

Originally posted by: Jeff7181
Duvie... exactly how is this inaccurate...

AMD is still the best bet when it comes to business/gaming/2D workstation either from a performance perspective or because of a superior price/performance ratio, while Intel offers the best in encoding and 3D rendering performance.

???

*EDIT*
I don't think many people here need to be told this. This is the way it's been for years... Intel for multi-media, AMD for gaming/business. That's the nature of the beasts. AMD's architecture is more suited for gaming and office type workstations roles... and Intel's architecture is more suited for multi-media and rendering workstations. And neither one performs that badly in the other's strengths that make it unusable. If the computer you're putting together will be used mainly for gaming, right now the smart choice is an AMD 64 system... if it's mainly used for multi-media type applications, the smart choice is an Intel system.

The thing is... most people either do a lot of video/audio encoding, or they do VERY little. Which IMHO makes an AMD 64 system a better fit for their needs.

Who said I was saying anandtech's conclusions were inaccurate...i said they have a lopsided suck-arse review....That is what I said..

Well, the tests performed clearly show that their conclusion is correct. Unless you're keeping score of benchmarks "won" by each manufacturer, the number of tests doesn't mean a whole lot.

I think this is a perfect example of why not to rely on just one source of info to draw conclusions.

 

[THUGSROOK]

face it guys ~ they all suck.

you wanna know how good something really is?
buy it and find out