OCAU thing I saw....

[Zebo]

Originally posted by: Duvie
I am sure many P4 fanboy ppl will love it...not so much for the numbers cause they dont relate to the words you were speaking, but for the comments themselves.....

You put in the big boys but negate them cause you think they cost too much...then you equate much costlier mid to high range chips versus the cheapest bottom of the line chip of the range..

Where the heck does the FX55 cost 1145???? I see it much cheaper at many sites....Damn talk about skewing the data...open up pricewatch once in awhile...It can be had at a reputale monarch for same price as the 3.73ghz P4.....and yes retail!!!


I would stop defending yourself as well....It are not doing a good job of it, frankly!!!

what he said. your review bleeds fanboy.

 

[friedrice]

Are you talking about the specific article or the web page as a whole?

 

[friedrice]

I'm interested in starting a hardware review page. Obviously I need a team of people for website development, graphics, and that sort. Also how do you get companies to send you samples down to test them? I'm really interest in either starting a new site, or improving an existing site. email me or reply if you are interested.

 

[Zebo]

Originally posted by: Duvie
how does the via board hampoer things as well???? Are not the NF 3 ultr and NF4 boards the best???

Least he ran ram at cas 3 instead of 4.

 

[Duvie]

Originally posted by: friedrice
I'm interested in starting a hardware review page. Obviously I need a team of people for website development, graphics, and that sort. Also how do you get companies to send you samples down to test them? I'm really interest in either starting a new site, or improving an existing site. email me or reply if you are interested.

I tried it once but dont have the time....My hats off to him for the time he spent...Just came a few laps short on the review, IMO...

I am talking about the review only....I have not seen the site....I can imagine it is nice!!!

 

[Zebo]

AMD Testing Platform:
Motherboard: Soltek K890Pro 939 (VIA K8T890 Pro chipset)
Memory: 1 GB OCZ DDR PC3200 3-2-2-8

How about CMD?

 

[friedrice]

CMD? Hey Zebo, do you work for a computer web page?

 

[DrMrLordX]

This is why I'll be waiting for some benchmarks of the 6xx series P4s from Anandtech.

 

[Zebo]

CMD is command rate either 1T or 2T in A64..makes huge difference 1T about 5-10% better in performance.

No I work for a DOD contractor..this stuff is my hobby I would'nt like it anymore if I had to devote any serious time to doing it.

 

[Zebo]

Originally posted by: DrMrLordX
This is why I'll be waiting for some benchmarks of the 6xx series P4s from Anandtech.

Long as Anand does them

 

[Duvie]

My stay at OCAU is limied...

Chainbolt in censoring out my comment son the review.....

He censored this comments and said they were inflammatory.....

QUOTE=chainbolt]CAS 3 is the Intel specified standard latency for the 925XE chipset. DDR2 you use with that chipset (as they did) is supposed to run at CAS 3 at 533 Mhz. Good DDR2 is running up to almost 700 MHz in 3-2-2-2 by the way. Explanations and test results here:

http://www.overclockers.com.au/article.php?id=340589

Their testing is correct because in order to show the "pure" CPU performance, you have to eliminate as much as possible any other influence. In many other tests they compared a A64 system running at high-end DDR2 in 2-2-2 with DDR2 running at 4-4-4. That is misleading if your intention is to compare 2 CPU with each other. Running both systems, A64 and P4 600, at CAS 3 is therefore the correct methodology.[/QUOTE]


WTH are you smoking?? The AMD board is DDR not DDR2....The correct methodology is to run the cpu on the best platform (which he did not) with industry standards for PC3200 (which he did not)...PC3200 is cas 2 in most all modules....Dont feed us that bunk. What a bunch of crap....Do you realize how much IO devices effect some of those scores yet how are you going to equalize them out over 2 platforms??? Dont neuter one cause the platform by Intel uses inferior latency ram at this point in time...Thi sskews the real results which is most ppl well build sytems with cas 2 or at least cas 2.5....Dont bother comparing AMD in this reviw if you will be doing a hatchet job.....

For you to somehow justify this is insane....

Go to Anandtech chainbolt and read the thread that called him out...listen to his responses...this is a piss poor biased review (maybe why you like it so much) thru and thru...

Why not include some mid range A64 processors in there??? Is it so he couldn't make his lame arse comparison to the bottom of the line A64 at a fraction of the cost of the p4 6xx series???

Why say the FX55 is 1145 dollars when it can be had as much as 200 dollars less then that and 150 less as aretail chip from a reputable site like monarch???

Why make absurd comments of p4 domination in audio encoding when it lost evry test except the last one where the older non prescott core won by a margin of 3%???

Why make absurd comments about video domination when even in his won benches the marghin of victory over the FX55 by either the 3.46 or 3.73ghz was around 3-5%??? especially since we all know had he not neutered th eram those 3-5% losses could have been less or even reversed to an AMD victory???


It was poor, and you supporting it shows your true color!!!

Not the only lame site I have seen recently!!!

I think you are mixing up 2 different things: the 3.46EE and the 3.73 EE have BOTH 2 MB cache and are both with a 266 MHz system bus. I'm actually also disappointed that the 3.73 is not doing better in comparison to a 3.46EE, but you can't claim that the larger cache has no impact, because these 2 processor have the same cache size.

The impact of the 2 MB 600 series versus the 1 MB 500 series is around 5% (depending on the application). That's similar to what I got when I reviewed the previous EE with 2 MB (Gallatin) with a 1 MB 500 series Prescott.

http://www.overclockers.com.au/article.php?id=331497

Given that you get the 2 MB L2 cache and all the other additional new functions that comes with the 600 series for almost the same price as the 500 series, it's not a bad deal - certainly not a breakthrough though.

Dont put word in my mouth...I didnt compare the 3.46 to the 3.73ghz in term of chace size??? that last line goes with the lines below it....I know that is a product of the longer pipeline of the prescott hurting the 3.73....One of the reason the prescott IMO is a poor move by Intel.....

If INtel would design nice new platforms around the stil viable northwoods we would see them still in the lead....I guess that is why we dont see it happening, huh???


Edit: ofcourse by lead I meant in the Intel camp!!!


Nonetheless he is a mod out of control!!! NO wonder that is such a lame site now..no discussions can be had....

 

[friedrice]

Well, anyone work on a computer web page that needs help? Or anyone interested in starting a new pc hardware web page?

 

[Zebo]

LOL so the moron is advocating underclocking (undertiming) A64 DDR ram and overclcoking (overtiming) DDR2 P4 ram. ..nonsence, you run ram at thier SPD's at least..certainly not cripple one while benefiting another.... D, tell the fool you can run DDR @ 1.5-2-2-8 1T and watch how A64 lays smack down.

 

[Duvie]

He has magically reappeared t2 of my 3 deleted post...possibly cause I said it was all being documented here and other could come look at what he is censoring......

Yeah he is a punk!!!

 

[Duvie]

And he is still talking about why neutering the AMD memory to cas 3 is the RIGHT thing to do....

and I quote!!!!

CHAINBOLT said,


In the same way as AMD has processors running at different bus speeds, so does Intel. You have Northwoods and even Prescotts running at 166 Mhz, 200 MHz, and now also at 266 MHz. The EE 3.46 and the EE 3.73 are both running at 266 Mhz. In order to test these processors you need a motherboard that supports a bus speed of 266 MHz. That is is currently only the 925XE chipset. Of course they also run at a 915 board, but in that case you would have to overclock these CPU. So why would you do this, when you have a 925XE board that supports boths types of CPU, the 600 series at 200 Mhz and the 600 series EE at 266 MHz, with the same board?

Detailed explanation how and why here:

http://www.overclockers.com.au/article.php?id=331497

In order to compare the sheer performance of 2 different CPU you have to eliminate as much as possible all other factors. For that purpose logic dictates that you run them as much as possible at the same memory latency. Running the A64 and the P4 system both at CAS 3 is reducing any difference you get to the impact of the processor, and that is what we want to see. If you use different latencies, the performance differences are "influenced" by the RAM, and that would be inappropriate in a CPU test.

A different question are the benefits of DDR2 versus DDR1. The better latency of high-end DDR1 running at 2-2-2 is supposed to be compensated by the higher frequency of DDR2 running up to 700 MHz at 3-2-2 and going up to 800 Mhz in 3-3-3. There are of course other factors as well, such as the pricing and availibilty. But DDR1 vs. DDR2 is not (and should not) be our point here, because you can run a 600 series P4 also with DDR1, by the way.

Chainbolt pauses to come up with some more lame arse reasons....

In reply I say.....

Why do you support running the DDR2 at some of the highest settings and not necessarily standard yet neuter DDR to cas 3 cause the DDR2 has poor latencies at this point??? Doesn't make sense...That isn't leveling the playing field is it???

CHAINBOLT said,

Because the purpose of this review was obviously to show the performance of the CPU, not of the memory. They did, by the way, neither use high-end DDR1 nor high-end DDR2. They simply ran both systems at the same CAS latency, and that is the correct way for testing CPU performance.


I like this one though....from his own mouth

The better latency of high-end DDR1 running at 2-2-2 is supposed to be compensated by the higher frequency of DDR2 running up to 700 MHz at 3-2-2 and going up to 800 Mhz in 3-3-3.

I said this contradicts his response to say that making the A64's cas 3 was making it more of a direct cpu comparison....

This guy is somthing else...No doubt elite there

 

[Zebo]

Ask why he's not running DDR A64 @ 267Mhz 3-2-2 if he really wants to make them "the same".. guy is full of shat.

I've never been to OCAU or "hardcoreware" now I see why

 

[lowlight]

I admit matching the memory latencies was the wrong thing to do, and I will never do it like that again.

From now on, I am going to set the ram to the 'highest performance that is readily available'. Basically use the best DDR you can buy for the AMD platforms, and the best DDR2 you can buy for the P4.

Like I said, I can take constructive criticism, and this was a good one.

 

[Duvie]

Originally posted by: lowlight
I admit matching the memory latencies was the wrong thing to do, and I will never do it like that again.

From now on, I am going to set the ram to the 'highest performance that is readily available'. Basically use the best DDR you can buy for the AMD platforms, and the best DDR2 you can buy for the P4.

Like I said, I can take constructive criticism, and this was a good one.

Well that will make you a better reviewer, no doubt....

I respect you tremendously for taking some of this heat.....see if newegg will send you a 3800+ and a 3500+ and just amend that into the test....wishful thinking???

I think you definitely go optimal.....If you don't have the best you still always run it at spd...Dont downclock anything....

the P4s issue of using DDR2 with it high latenoies right now is their choice. DDR2 will get better in time and be faster. Unfortunately perhaps by then AMD will up its memory controller as well....Too tough to try to isolate cpus nowadays...

 

[Zebo]

Good call. Hey I liked your writing style and believe your benches are sound (i've tested a couple to find error) Just think you got caught up in the 2mb cache hype, thinking it would do what it did for northy (it did'nt by a long shot) and your invections permeate that belief even though your own benches speak otherwise.

 

[lowlight]

The trouble with doing as many benchmarks I do, is that I feel the need to come up with something to say for each one. And in a review of a CPU where the only real new thing that COULD affect performance is the cache, that's all I really had to talk about.

However the most important part of this launch are the Halt State and Speedstep. And I was one of only 2 sites to put that to the test. Check out my results on page 11 and you'll agree that they are quite impressive (of course only because the Prescott was so bad to begin with for heat, but hey this is a nice step up).

 

[Duvie]

How much does it cost t buy that worldbench test kit???

 

[lowlight]

$250

http://www.pcworld.com/resource/article/0,aid,116892,00.asp

 

[Zebo]

What did OCAU Chainbolt have to say about running FX-55@ 10 x 260 3-2-2-8 1t? fair is fair. Just get some VX or Bh-5 on DFI NF4 can do this trivially.. Actually can do 1.5-2-2 but we don't want to put pressys at a disadvantage....

 

[Duvie]

Originally posted by: Zebo
What did OCAU Chainbolt have to say about running FX-55@ 10 x 260 3-2-2-8 1t? fair is fair. Just get some VX or Bh-5 on DFI NF4 can do this trivially.. Actually can do 1.5-2-2 but we don't want to put pressys at a disadvantage....

He responded by saying it was ran at 3,2,2,8 and bypassed the whole 267fsb setting for 533ddr....So no answer....

 

[Duvie]

Duvie asked:

Using your logic when P4c's were compared to Athlon xps should we have reviewed them in single channel mode??? I mean the P4 needed the bandwidth because it had been greatly bottlenecked with single channel, however single channel amd xps were in perfect balance.....

Chainbolt replied....


Actually they should have. And I am also not dismissing the argument that a "real world" scenario could have been tested with good logic as well, in terms of running both systems at the "best of their capabilities". That would have been a A64 with DDR1 at 2-2-2 and a P4 600 with DDR2 at 3-2-2. But then, as I said, to do justice to the DDR2 system, you would have to max it out so to show what DDR2 can do, and that would have been running such a system at FSB 350 Mhz in 1:1 (DDR 700) or even with the memory at 800 MHz. Also keep in mind: You can run a P4 600 series (not the EE though) at dual channel with the same DDR1 you use for a A64.

He has been starting to agree with my logic in the arguement but the last few lines shows you he is pretty much only in the favor of putting the p4 in the advantage....Now he wants to jack up the fsb to 350...Nowhere does he say running the A64 at anything better then cas 3 still and pc3200 still.....

he knows damn well the P4 thrives on bandwidth more so and that DDR2 latencies are high right now...he also knows AMD likes timings and doesn't need all the bandwidth it has so much of now.....He is just picking the path that favors his side and his architecture...
he is so blind to see it is those differences in architecture that make trying to put this test into parity biased within itself....