prescott 3,4ghz benches

[THUGSROOK]

Originally posted by: Duvie
I fail to see where prescott was an upgrade..Please show me!!!

SuperPI

and we all know how important that is

 

[LTC8K6]

Show you what, Duvie? Intel has had the statement about the case specs with their retail processors for a long time now. It's a yelow sheet.

Lots of folks aren't having any trouble with their Prescotts, though.

I am happy with mine.

Like everything else, you only tend to hear from people having trouble.

Like with those SLI boards.

 

[Sc4freak]

Throttling? As I have experienced, the throttling on Prescotts is more frequent at higher voltages and clockspeed. This is not because the heat generation rises so much that it forces it to throttle, it is because the chip seems to lower the threshold for throttling at higher clockspeeds and especially higher voltages. If you set it to default voltage (1.375v) and keep it under 70C, it won't throttle.

Edit: And my performance raises linearly with my overclock, I definately get higher performance when at 4.0ghz compared to 3.4ghz. No idea whats wrong with your chip.

 

[dev0lution]

I got similar results running at 3.6 (from 3.4) in benchmarks. But then again I can't really mess with anything other than a piddly FSB increase and memory timings with this intel board.

 

[THUGSROOK]

ok fine....

prescott owners defend their prescotts.
any AMD owners wanna defend the prescott?

i didnt think so.

 

[Sc4freak]

Uh-huh. When was I defending the Prescott? I was merely stating the facts that I discovered from my firsthand experience. I never said that the Prescott was better, or the Prescott was cooler etc.

 

[LTC8K6]

Defend Prescott from what?

The majority of Prescott owners aren't having these problems. that tells me there's probably nothing to defend Prescott over.

Tom had no real problems with both AMD and Intel systems during the stress test. No throttling for the 3.6Ghz "Presshot" either. Stock cooler and TIM pad.

"Monday, Jan. 10 2005: We are done today! Both test systems survived the start into the the new year without problems. After system crashes in the beginning, both systems achieved a flawless running time of 11 days and 22:53 hours for AMD and 17 days and 14:28 hours for Intel."

If Tom can do it properly, I sure as hell can!

When I build my A64 system to partner my "Presshot", I am sure I'll have the same amount of trouble as I had with the Intel system. None.

 

[AristoV300]

I have never had any problems with my Pressy. The stock cooler is crap but with my XP-90 temps never exceed 50C under full load. Never had any throttling problems either. I am much happier with the overclocking potential compared to my prior NW's.

 

[THUGSROOK]

Originally posted by: LTC8K6
Tom had no real problems....

who is Tom?

 

[THUGSROOK]

Originally posted by: AristoV300
The stock cooler is crap but with my XP-90 temps never exceed 50C under full load.

asus moboards dont have a cpu temp

your 50*C is the PWM temp ~ thats freakin hot!

 

[clarkey01]

Originally posted by: LTC8K6
Defend Prescott from what?

The majority of Prescott owners aren't having these problems. that tells me there's probably nothing to defend Prescott over.

Tom had no real problems with both AMD and Intel systems during the stress test. No throttling for the 3.6Ghz "Presshot" either. Stock cooler and TIM pad.

"Monday, Jan. 10 2005: We are done today! Both test systems survived the start into the the new year without problems. After system crashes in the beginning, both systems achieved a flawless running time of 11 days and 22:53 hours for AMD and 17 days and 14:28 hours for Intel."

If Tom can do it properly, I sure as hell can!

When I build my A64 system to partner my "Presshot", I am sure I'll have the same amount of trouble as I had with the Intel system. None.

The majority of Prescott owners dont know they have one (dell buyers). Not many enthusiasts have got one. If the PC world was more awake and informed about heat, slower clock for clock, power consumption they'd ask for a northwood, or even a walk to the darkside. If you remember the preshott was delayed 6 months because the "thermal target" was adjusted in thier words, lol.:Q

 

[clarkey01]

Originally posted by: Duvie

Originally posted by: LTC8K6
Intel has a list of chassis specs and approved case designs for these processors. but few people pay any attention to it.

Thermally Advantaged Chassis - no not BTX

i am sorry but that is a bunch of bunk...Intel supposedly comes out with a newer revised core in the prescot and now we have to change the whole system cause Intel can't get there act together and quit producing these power hog, thermal nuclear devices!!!!

Then give us a product that is:

1) worse clock for clock then its predecessor. No need for a revision as far as I have seen, considering the lame amount they have been able to ramp them up. They hobble it with a 30 stage pipeline to be able to ramp it higher and this is the best they can do...Oh trouble!!!

2) warmer per clock: Have to have better case cooling, stock cooling out of the box is ineffective unless you have like 12 case fans so you have to buy aftermarket cooler....BS!!!! I guess Intel should spec water blocks and force us to get water coolers...LOL!!!

3) More power consumption. Unprecedented amounts of power needed...

I fail to see where prescott was an upgrade..Please show me!!!

Very True. But only the enthusiasts know that. Sad but true

 

[Sc4freak]

Originally posted by: THUGSROOK

Originally posted by: AristoV300
The stock cooler is crap but with my XP-90 temps never exceed 50C under full load.

asus moboards dont have a cpu temp

your 50*C is the PWM temp ~ thats freakin hot!

They don't? Then whats the "PWM temp"? As far as I know, PWM stands for Pulse Width Modulation, and is used to regulate fan speed. Care to explain?

 

[THUGSROOK]

PWM = PoWer Management

 

[THUGSROOK]

by comparison my abit never reaches higher then 36*C PWM

and yes the asus PWM and abit PWM are calibrated to within 1*C of each other ~ its all been tested long ago by ppl that have been around a very long time

 

[THUGSROOK]

lets see now.... 50*C PWM would put the CPU temp at close to 76*C (roughly)

nice of asus to trick you like that huh? yea....

 

[THUGSROOK]

oh BTW one more thing ~ p4 doesnt have an ondie temp sensor either.
so there is no way to know for sure what the cpu temp really is.

lesson over

 

[LTC8K6]

Tom's Hardware Guide, a well known web site Thugs. The P4 3.6 ran for ~17 days at 100% load in a case with the stock cooler. Earlier, Tom's had bashed that chip for being too hot if you used poor quality thermal compound.

P4's have always had on die temp sensors as far as I know. It's what allowed them to keep from dying if the heat sink failed. It was one advantage they had over AMD chips for a long time. An external sensor, used in AMD systems, was not accurate or fast enough to save the chip in the event of heatsink failure.

PWM is Pulse Width Modulation and it refers to a power regulation and supply chip on the motherboard.

 

[TStep]

LTC8K6:
I didn't follow the THG testing at all, so I'm at a loss as to the content. Were any measurable benchmarks performed ensuring the 3.6 is operating at 3.6 for the most part? Was the throttling monitored? Was there any throttling?

 

[clarkey01]

Originally posted by: TStep
LTC8K6:
I didn't follow the THG testing at all, so I'm at a loss as to the content. Were any measurable benchmarks performed ensuring the 3.6 is operating at 3.6 for the most part? Was the throttling monitored? Was there any throttling?

No worries, Not a very credible source (toms).

It was Zebo or Dapunisher were talking about the tests and one of them pointed out an inaccuracy in voltage/temps while this test was going on , think it was in a thread about the athlon 64 running hotter then a Prescott?

 

[Zebo]

Lots of folks aren't having any trouble with their Prescotts, though.

How would they know?

It's not like Prescotts, jump out and say -- "Hey dude, I'm only on 80% of the time while your playing EQ2 after 15 min, sorry I just get too hot above 3.2" -- Sound testing, at varying freq, similar to what OP did is the only way. (though I still not satisfied with his limited test suite or efforts to keep mem speeds same)

They don't know, w/o thourough testing because they would never notice...stevty2889 stumbled upon the throttle pheonmina in our chinebench thread. The OP did so while benching HL at varied freqs. I would wager most user are throttling and are just clueless about it's effects.

Also many Prescott buyers are wise and buy xp120's, which is bascially poor mans water cooling in my experiance, and probably keeps chip from doing so at stock speeds.


Bottom line, more data reqd. Toms does'nt count. He's worthless as I discoved once when I had same config as him and benched almost double in all games. HE had "PSU failure" 2x in prescott test, and instead of blaming power hog he chose to dismiss it as Antecs fault. His methodolgy is screwy, he never cites it, and says the northwoood 3.2 uses only 52W loaded when intels own TDP is 82W and they are know to be conservative. Basically if you quote Toms you loose all credibiltiy IMO.

 

[clarkey01]

lol - there ya go .

Clarkey Tip for you TStep: AVOID TOMS

 

[Duvie]

Originally posted by: THUGSROOK
oh BTW one more thing ~ p4 doesnt have an ondie temp sensor either.
so there is no way to know for sure what the cpu temp really is.

lesson over

 

[Duvie]

Originally posted by: LTC8K6
Tom's Hardware Guide, a well known web site Thugs. The P4 3.6 ran for ~17 days at 100% load in a case with the stock cooler. Earlier, Tom's had bashed that chip for being too hot if you used poor quality thermal compound.

P4's have always had on die temp sensors as far as I know. It's what allowed them to keep from dying if the heat sink failed. It was one advantage they had over AMD chips for a long time. An external sensor, used in AMD systems, was not accurate or fast enough to save the chip in the event of heatsink failure.

PWM is Pulse Width Modulation and it refers to a power regulation and supply chip on the motherboard.

The guy was a moron....he had temps off for some reason and they mysteriously would jump with no explanation. His cpu and chipset fans were backwards. For the first week his INtel systm was hetaing the amd system and couldn't figiure the logic of this, until ppl on the net had to tell him. He used open case for awhile which is a not a true measurement
of most ppls systems....Yeah no credibility. It was like a noob doing the test....I dont trust his methodology or that he didn't go into it with a set goal to squelch other sites claims the prescott was throttling a certain amount of the time even at stock....

NIce below freezing temps of Germany in the winter with I am sure a nice climate controlled lab....Again not indicative of most around the world who do not have AC...

Again wait until the summer rolls around...

It also doesn't help these damn 6800's run in some instances up to 68-75c on their own. That is tremendous case loads of heat...Was Tom running one of these???


PWM on my Abit mobo was a temp sensor that was located around the area of the power caps and power regulation of the board...It was usally 41c. If it went above 45c I usually had issues. My cpu was usually about 15-20 over that....

 

[TStep]

Thx Clarky, but TStep (as blackinches would say in his 3rd person voice) isn't a young whippersnapper. I've found if you can sift through the sea of FUD, you'll find the answers your looking for. Be it Toms FUD, forum FUD, whatever; it all has meaningful info in there somewhere.

As I see it, with dual cores around the corner, heat management will be of significant concern for the enthusiast regardless of the platform. The sheep will just follow the heard, and take whatever they get and not give a crap. Be it user engaged cool-n-quiet or invisible to the user clock throttling, both platforms are providing means of smart cooling management now, and I could only assume smarter, and perhaps more devious, methods in the near future. stevty2889 and Strawa provided the first meaningful info that I've come across on what I would consider a devious solution -- invisibly engaged and detrimental to the performance the owner thinks he/she purchased. An effective solution, would be somewhat akin to as how ATI handles trilinear filtering -- invisibilty, and in essence, the user sees no reduction in performance or quality.

Now I believe Dothan manipulates power to large blocks of cache to reduce power requirements if the cpu determines them unnecessary. If that's effective and not detrimental to performance, great, it's a step in the right direction. By no means am I an Intel basher (w/ no less than 1 C, 1 EE, 1 Mobile, and 1 Willy setting on the shelf) but if stevty2889's and Strawa's results are an indication of how current and future cpus are going to perform, it would be nice for one of the major review sites with some horsepower to get an Intel rep on the hotseat for an explaination similar to the poor sap from ATI who had to answer for adaptive trilinear.

signing off >>Joan of Arc