Quiet High-end P4 3.4C - is this possible? Your help please

Got a shopping cart of parts that you want opinions on? Get advice from members on your planned or existing system (or upgrade).

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

shathal
Posts: 1083
Joined: Wed Apr 14, 2004 11:36 am
Location: Reading, UK

Post by shathal » Thu Apr 22, 2004 10:01 am

Ralph suggested cooling my 3.4 GHz with the ThermalRight SP-94 - I took him by his word. Might have been more necessary with me, since I'm aiming also for a "toast-a-lot"(tm) high-end next-gen graphics card.

Should hopefully be able to report soon on the noise-level of the setup as is :).

burcakb
Posts: 1443
Joined: Tue Mar 09, 2004 9:05 am
Location: Turkey

Post by burcakb » Thu Apr 22, 2004 12:57 pm

Since you're in UK, you might want to look at Noiseblocker products. Currently I'm viewing those and I'll try and get some 120 and 80 mm fans. They seem to push a lot of air with minimal noise. From what I could make out with my limited german, most people compare the 120mm to the Papst without the click.

If I could say money was of no concern, I'd be looking at the SilentMaxx proSilence 503W semi-fanless PSU. It's a muchowatt PSU that runs fanless until it gets to 65C then starts a quiet fan to drop it down to 45C. Neat but expensive.

I haven't seen a high-end Dell yet, but the lowish-end (2.4GHz) ones had a tight custombuilt duct that ran from the CPU HSF right up to the 92mm fan. That helps a lot. Plus their cases contain noise very well. If anyone here were to make such a goodfitting duct, I'm sure the Dell would sound like a rock band.

Putz
Posts: 368
Joined: Thu Aug 21, 2003 1:25 am
Location: Ottawa, Canada
Contact:

Re: Quiet High-end P4 3.4C - is this possible? Your help pl

Post by Putz » Thu Apr 22, 2004 1:27 pm

I know you said this:
shathal wrote: What I DON'T want/need:
- Modding the case.
- Glass panels in case.
I am quite happy with just a "solid box", really. Long as it ideally looks nifty, but is quiet. I'm not a modder junkie (yet), so it's not a concern :).
but really, for a quiet system, the restrictive fan grille on the back of that case has got to go.

shathal
Posts: 1083
Joined: Wed Apr 14, 2004 11:36 am
Location: Reading, UK

Post by shathal » Thu Apr 22, 2004 1:43 pm

I did post this as my first post on this forum.

Having read around here, I've come to understand that likely I'll not get overly far without a little bit of snipping and modding.

I'll first see how it goes though, compared to my current system. Sound is all about relative anyway.

If I can be lazy, I'll try to not chop things up. But the forums are slowly dawning on me and ... well - shall see. I'm in need of a PSU as well, so will see how things go.

One of my main arguments for the NoiseTaker 475 (apart from MikeC's fantastic review) is the seperate 12V rail for the CPU-feed and the usual 12V. Plus, high quality power ...

I'll have a look at the SilentMaxx PSU though - thanks for the hint.

(Only components I need now really, PSU + Graphics card ... the DVD-stuff too, but that's less critical for now. I can recycle existing stuff for now :D)

sthayashi
*Lifetime Patron*
Posts: 3214
Joined: Wed Nov 12, 2003 10:06 am
Location: Pittsburgh, PA

Post by sthayashi » Thu Apr 22, 2004 1:52 pm

Cutting a fan grill is really really easy with Tin Snips. There's a mention of this in the Cases & Modding Sticky by Crisppy. Using his advice, I purchased a set of Wiss Tin Snips and quite easily cut out the grill on the back of my SN41G2 Shuttle. Since it's basically the first (and only) case mod I've done, I kept the cut out grill. It still sits on my kitchen table because I think it's too cool to just throw out. :D (this alone should tell you how n00b I am when it comes to case modding).

Anyways, back to the point, I think you'll find it just as easy as I did. Basically, cutting it out was about as difficult as cutting heavy construction paper with a set of good scissors.

shathal
Posts: 1083
Joined: Wed Apr 14, 2004 11:36 am
Location: Reading, UK

Post by shathal » Thu Apr 22, 2004 2:11 pm

Oh yeah, I came across that post. I've read, and learned my lesson to buy WISS snips :).

One of the forums even mentions places where I can get them in the UK, so I am armed with information. Now, to see if / how much chopping will be necessary :).

Most of the system will be uniquely silent what with the distinct lack of a PSU. Deathly silent, nigh, I am tempted to say :D

roaf85

Post by roaf85 » Fri Apr 23, 2004 10:08 am

Ok someone pointed out that dells hit a thermal throttling and there are reasons to make me believe this and there are reasons that might not. First off I have not extensively prime tested a Dell dimension 8300. However from some 3dmark01 tests I have saw some Dell still score within a good acceptable performance range. However, the overall responsiveness of most dell systems I have seen tend to be sluggish. Now you can get the same config in a Dell Dimension 8300 for about 400 dollars less the XPS. Granted the XPS has better support but it also has 4 more fans. So are we to assume that the 8300 does hit a thermal throttling point? I couldnt say without testing. From ARM testing a fully loaded Dimension would hit 40dbs plus on one fan?

I could type all day on this topic. I have before in another thread I did. However to get a clearer understand I am wondering if I could chat with Steve at ARM systems. I know he is a busy man and I am not really serious to buy right now but in the near future I will. Let me know in a PM. Thanks.

shathal
Posts: 1083
Joined: Wed Apr 14, 2004 11:36 am
Location: Reading, UK

Post by shathal » Fri Apr 23, 2004 11:46 am

....? Might be wrong topic ...? :)

Anyway.

And thus, on the eve of the weekend, he sat down and began ... THE ASSAULT! :twisted:

... but - oh dear - it was but a feint. :oops:

With my flatmate out of the house for a few weeks, I had a clean conscience in setting my system together 90% of the way.

... but ...

... looking at the accousti-pack I'd received, and looking at my chassis, I was wondering "THAT's supposed to fit?". Looking closer at the instructions for the dampening kit and ... oh dear ...

Turned out that while QiuetPC did send the right chassis (C6607) they send me the wrong dampening kit (C6606), which is completely incompatible, pretty much.

So, I've e-mailed them, but have to wait until at least tuesday, to receive the kit. Darn. This will leave me for this weekend with trying to resurrect my XP Partition, which seems to have suicided itself rather unpleasantly.

Observations on the chassis
It's a decent chassis, all in all.

The fan-holders can't cope with the Evercool 120 mm aluminium fans (WAY too big), so I'll need to get some gromets and ponder on how to attach them (will require longer screws one way or another, especially since I'll "reverse" mount one of them to the front of the chassis.

I'll likely snip off the rear fan-guard, it does seem rather obstructive and could have been designed better. I will do so, however, only once I've found a good wire-guard that I can put there instead.

Any good ways for me to buy one of those, or what's the froogle / kelkoo keyword for me to search for on these? I'm talking about the minimalistic wire-based finger guards which are circular in pattern.

So ... disappointing, all in all :(. Hoped to be down a lot further today. Oh well ... I've got a little shopping list for stuff to get for tomorrow, and will see how that goes.

ThermalRight onto motherboard => the X-plate is in place, further assemblage is pending putting the mobo into the chassis, as I do *NOT* want to be putting the weight of the ThermalRight onto the mobo while it's unmounted.

... oh yeah, and otherwise, fitter the I/O shield. Bit of a tighter fit that I'm used to, but worked in the end.

So ... pitty, really. The thermal pads being the wrong ones cripples me in the starting round so far... :(

Ralf Hutter
SPCR Reviewer
Posts: 8636
Joined: Sat Nov 23, 2002 6:33 am
Location: Sunny SoCal

Post by Ralf Hutter » Fri Apr 23, 2004 1:28 pm

roaf85 wrote:Ok someone pointed out that dells hit a thermal throttling
It wasn't me.

AFAICT, Dells do not thermally throttle, they just run warmer than we're used to doing. There's a big difference.

They are designed by very competent engineers to run a certain way, and I'll guarantee that doesn't include running so hot that they throttle, at least not under any normal circumstances. Re-read the threads I linked you to, there's good info in them.

shathal
Posts: 1083
Joined: Wed Apr 14, 2004 11:36 am
Location: Reading, UK

Post by shathal » Fri Apr 23, 2004 1:48 pm

(off topic - albeit my own one at that...)

One thing that makes no sense to me in that peculiar design choice of Dell - running the internal temps so "high", it MUST affect their RMA-rates (HD's dying, and so on).

Ah well - suppose, at the end of the day it's more about total P&L rather than primarily reliability.

While "the big guys" tend to talk about reliability, I tend to find that the concepts of "margin" (which is really what being a corporation is about) and its friends tend to go somewhat against that.

Not saying you can't get reliable systems. Not at all.

Just am of the opinion it's usually easier to get them the DIY / "pick & chose" way.

Ralf Hutter
SPCR Reviewer
Posts: 8636
Joined: Sat Nov 23, 2002 6:33 am
Location: Sunny SoCal

Post by Ralf Hutter » Fri Apr 23, 2004 2:13 pm

shathal wrote:(off topic - albeit my own one at that...)

One thing that makes no sense to me in that peculiar design choice of Dell - running the internal temps so "high", it MUST affect their RMA-rates (HD's dying, and so on).
The HDDs are mounted up front, right behind the bezel so they get the coolest air on it's way into the case. I've checked the temps with Dtemp and they generally stay in the high 30°C to mid 40°C range, which should be OK.

shathal
Posts: 1083
Joined: Wed Apr 14, 2004 11:36 am
Location: Reading, UK

Post by shathal » Sat Apr 24, 2004 2:07 am

Further updates from the front:
After last night's disappointing finding, a new day brings a needed calm perspective on things.

Findings on the chassis:
Still overall quite happy with it. Some of the machisms there were quite ... "interesting".

For instance, the manual for the chassis was INSIDE the case. That on its own would have been OK, were it not for the fact that the side-panel swivels to the SIDE (!), and doesn't get pulled back, as per the more traditional chassis.

That cost me a few minutes to figure out (along with the locking mechanism), but I've likely not been at my intellectual best yesturday (is my excuse at any rate).

The fan-holders that were provided for the 120mm fans had to go, as there's no way that they'll hold the massive Evercool aluminium frame, so off they went.

The rear-fan grill will likely be going "very soon", but I'd need to replace it with a finger-guard. Can anyone give me google-able keywords to search for, so I can get ahold of the circular-pattern wire-based guards? Thanks. The grill is also dented a little, which would be more an optical thing, but since it'll be going once I've ordered my WISS snips, it's not on my priorities.

EDIT:I've found the term, it's incidentally "Metal wire grill" - voila :)

"SINGLE-SIDE mounting" - all the 5 1/4" drive-bays only get screwed in on the side where the side-panel comes off. While this is rather "nice" for DVD-ROM's and co (i.e., devices which takea full lenght slot), it does make installing my Zalman Fancontroller somewhat of a pain (so far, only been able to fasten it with a SINGLE screw). Shall try to get some longer screws and washers/nuts ... see if I can improv something.

Failing that, I'd need to see if/how I can figure out to get a "full length" plate to attach to it. Little irritating, but oh well.

Spirits still high :).

On the shopping list for today:
====================
- Screws (longer), for Zalman Fancontroller (+ nuts + washers) and the Evercool face.
- "inverted Screwdriver to tighten the ThermalRight SP-94 correctly. I seem to have lost mine somehow...
- 1 (one) brass stud for fixing mobo to chassis. Can't find any spare studs at the moment, and somehow the Accousti chassis came short of one stud for me to mount my full-size ATX board. Not a biggie - I just tend to prefer to fix a mobo in ALL locations.

Only the paranoid survive and all that :).

Ralf Hutter
SPCR Reviewer
Posts: 8636
Joined: Sat Nov 23, 2002 6:33 am
Location: Sunny SoCal

Post by Ralf Hutter » Sat Apr 24, 2004 5:28 am

shathal wrote:Further updates from the front:For instance, the manual for the chassis was INSIDE the case. That on its own would have been OK, were it not for the fact that the side-panel swivels to the SIDE (!), and doesn't get pulled back, as per the more traditional chassis.
Ha! Once I got a case with a lockable door and the keys were in a box taped to the floor of the case. Had to bend a coat-hanger and maneuver it through the open 5.25" bay to dislodge the box so I could shake it towards the front of the case to retreive the key!
shathal wrote:The fan-holders that were provided for the 120mm fans had to go, as there's no way that they'll hold the massive Evercool aluminium frame, so off they went.
SOP. If you look at my reviews of the SLK3700's, you'll see that I say "pull off those silly purtple fan holders and throw them as far as possible. Note that you'll probably have to drill your own set of fan mount holes because, unlees they've made changes to the 6607, the hole pattern for the plastic mounts isn't the same size as the hole pattern for a 120mm fan.
shathal wrote:The rear-fan grill will likely be going "very soon", but I'd need to replace it with a finger-guard.
I can't believe that you'd have any trouble finding a fan grill over there. They're a dime-a-dozen over here. They're available at even the crappiest retail sotores, and at every online store as well. I've got boxes full of them, if I was strong enough, I'd frisbee a few over your way.
shathal wrote:"SINGLE-SIDE mounting" - all the 5 1/4" drive-bays only get screwed in on the side where the side-panel comes off. While this is rather "nice" for DVD-ROM's and co (i.e., devices which takea full lenght slot), it does make installing my Zalman Fancontroller somewhat of a pain (so far, only been able to fasten it with a SINGLE screw). Shall try to get some longer screws and washers/nuts ... see if I can improv something.
Isn't that a PITA? I don't know why Acousticase would source that bastardized version of a 6A19 to use as the basis of their new case. Surely the "real" 6A19 with drive rails on both sides of the bay wouldn't cost them that much more. Oh well, bean counters...


shathal wrote:On the shopping list for today:
====================

- "inverted Screwdriver to tighten the ThermalRight SP-94 correctly. I seem to have lost mine somehow...
"nut driver"

shathal wrote:Only the paranoid survive and all that :).
As my granny used to say "overcoming adversity builds character". :)

Putz
Posts: 368
Joined: Thu Aug 21, 2003 1:25 am
Location: Ottawa, Canada
Contact:

Post by Putz » Sat Apr 24, 2004 5:40 am

Wire fan grilles are available for cheap at a lot of places. You shouldn't have any trouble finding them.

The stamped grille for the rear fan on your case is not "dented" just for aesthetics -- that also is supposed to reduce turbulence. However, given the very restrictive pattern of that particular grille, turbulence and airflow impedance will both be very difficult to eliminate, "dent" or no. Glad you are not opposed to ridding yourself of it.

Sounds like you're having fun! What you are experiencing, in my experience, tends to follow the pattern for the average computer build. Slight hurdles, but they just keep life interesting! Keep those spirits high! It will be worth it in the end.

(Wait... if you read SPCR regularly, you'll know that there IS NO END! I guess it's one of those "all about the journey, not the destination" deals.)

shathal
Posts: 1083
Joined: Wed Apr 14, 2004 11:36 am
Location: Reading, UK

Post by shathal » Sat Apr 24, 2004 6:41 am

Right - more updates, mixed blessings and frustrations.

1 - got myself some 3mm x 30 mm screws (the ONLY length they had over 12 mm *ARRRH) + nut screws. Along, have gotten myself a set of 3.5 mm, 4 mm, 5 mm and 6 mm washers.

You'll read shortly WHY I wanted those washers.

2 - Got myself a set of screws for ca. 20 mm length for the Evercool 120 mm fans.

3 - Got myself (again) a nut driver. Haven't found the old one yet, which was actually what I'd expected, coming home (what with sod's law and all).

Successes / failures - The Zalman fancontroller:
Fitted the Zalman Multifan controller better. I needed the set of washers, to do get a little more "hold" on the chassis for the 30 mm screw.

]---[[[[---||-[]--------

Where, increasing size '[' symbols stand for the different size washers. These are, 3.5 mm, 4 mm, 5mm and finally a 6 mm washer. (had to go large). Was conteplating getting a 7 mm washer as well, but got tired of things. This works - for now :).

The || stands for the Zalman's plate (which incidentally bends out of shape REAL easy... :evil: ).
And the [] stands for the nut at the end (overkill mayhap, but since the nuts came with the pack, I might as well.

I had to do a similar setup on the other side, to try and get a little solidity for the Zalman. It holds in place now - mostly.

Part of the problem is that it's at a slight angle. Will see how much of a problem that is, when I shove DVD-drive(s) into it.

Successes / failures - The 120 mm fans.:
A lot of frustration on this front. The "easiest" bit was the cutting up of the grommets I got (as I couldn't squeeze them into the hole, I simply cut them into two, intending to squeeze them at each contact of the chassis with the screw/fan :).

My first finding was that the screws I got (helpfully) DID fit through the holes (barely). My other pleasant surprise was that the 6607's holes lined up exactly with those of the Evercool fans. :)

So I mounted both. To my surprise, while the rear fan was "somewhat stable", the front fan didn't seem to have the very same screws biting into it. The next size up ... that'd force me to increase the holes. Without a drill at the moment, something that'd be less than fun.

Following half an hour of frustration, contemplating on what to do, I've conveniently remembered something that SPCR's own BLUEFRONT had done.

Since I'd bought a pack of 100 cable ties as well, I recalled his experiment of using them to attach a fan (albeit I don't know why or how he came to that). Result?

FANTASTIC :D.

The fans are pretty tight in place, and the grommets should help things too.

Also replaced the last missing motherboard screw (they only sold packs of 50. Anyone need 49 motherboard screws?), so that's good :). Will likely have to do some cutting on the accousti-kit (once I get the right one), due to the massive size of the rear fan. Shall see what I come up with. If need be, I've got a bunch of spare cable ties anyway.

So ... we're getting somewhere. Mixed successes, but overcoming things so far... :).

EDIT:
Corrected a few things.

Found also the wire-fan grills, as per this place.

http://www.dealtime.co.uk/xPC-Antec_ANT ... EACH_10PKS

There's shops charging 7 quid for those? Feh ... it's just a bit of wire.

Anyway, the fan grill is on the backburner anyway, as I'm not sure it'd be held in place with my plagiarized Cabletie retention system :). Shall see... :)

shathal
Posts: 1083
Joined: Wed Apr 14, 2004 11:36 am
Location: Reading, UK

Post by shathal » Sun Apr 25, 2004 5:30 am

DVD-Writer is ordered, as per SomtimesWarrior's recommendation, it's an NEC NEC 2500a (see thread at: http://forums.silentpcreview.com/viewtopic.php?t=12146).

Now, on to finding a nice, quiet, broad format-supporting and black-bezeled DVD-ROM for main use... :)

roaf85

Post by roaf85 » Sun Apr 25, 2004 7:23 pm

Ralf Hutter wrote:
roaf85 wrote:Ok someone pointed out that dells hit a thermal throttling
It wasn't me.

AFAICT, Dells do not thermally throttle, they just run warmer than we're used to doing. There's a big difference.

They are designed by very competent engineers to run a certain way, and I'll guarantee that doesn't include running so hot that they throttle, at least not under any normal circumstances. Re-read the threads I linked you to, there's good info in them.
Ok Ralf I guess I have some questions here. First off you say they are designed by very competent engineers to run a certain way? I do not mean to be rude but all dells use is a plastic duct or whatever you want to call it over a big heatsink that looks a lot like something from thermal right with probably some thermal paste of some sort. They do the same wtih processors up the 3.4 ghz. Now why couldnt someone cool a 3.4 ghz P4 with a duct and a 120 mm fan? I mean to be honest if they are not thermal throttling why worry? I mean there is no data that proves a processor running at 50 C will last longer than one running at 40 C. Is the only way to cool a 3.4 ghz processor quietly is to run a zalman 7000 at 12 volts with a 12 volt case fan?

All I am trying to figure out here is if I can not build a as quiet as dell dimension 8300 computer using a 3.4ghz processor cooled but a zalman 7000 in a Antec Sonata case with the 380watt true ps and one rear exhaust fan, with a ATI 9800 XT compared to the same specs on the dell then I guess why bother?

JohnMK
Posts: 159
Joined: Sun Aug 25, 2002 7:37 pm
Location: Seattle
Contact:

Post by JohnMK » Sun Apr 25, 2004 7:46 pm

My 3.4GHz P4 runs fine with a Zalman 7000 CNPS at 5 volts. I'm using an ARM Stealth kit case/psu. I do video encoding work and pay close attention to my P4's temp at full load, after a few hours it tends to be in the low 50's.

wumpus
Patron of SPCR
Posts: 946
Joined: Sat Sep 06, 2003 9:57 pm
Location: Berkeley, CA, USA
Contact:

Post by wumpus » Sun Apr 25, 2004 11:52 pm

I think you may have the coolest running P4 3.4 in the history of recorded time. Particularly with ANY heatsink and a 5v fan.

shathal
Posts: 1083
Joined: Wed Apr 14, 2004 11:36 am
Location: Reading, UK

Post by shathal » Mon Apr 26, 2004 5:15 am

Oh, shall see :).

I hope to have temps like that myself too :).

wumpus
Patron of SPCR
Posts: 946
Joined: Sat Sep 06, 2003 9:57 pm
Location: Berkeley, CA, USA
Contact:

Post by wumpus » Mon Apr 26, 2004 6:43 am

Basically I'm saying I don't believe him. Just look at any of SPCR's reviews of even slower processors running with a 5v fan and the Zalman..

50c is a "best case" load scenario for a processor HALF his speed. (1.8 vs 3.0) Meaning, he's generating a ton of heat beyond what this test would show. 1.8 is so quaint..

http://www.silentpcreview.com/article57-page5.html

Rusty075
SPCR Reviewer
Posts: 4000
Joined: Sun Aug 11, 2002 3:26 pm
Location: Phoenix, AZ
Contact:

Post by Rusty075 » Mon Apr 26, 2004 7:22 am

wumpus wrote:Basically I'm saying I don't believe him. Just look at any of SPCR's reviews of even slower processors running with a 5v fan and the Zalman..

50c is a "best case" load scenario for a processor HALF his speed. (1.8 vs 3.0) Meaning, he's generating a ton of heat beyond what this test would show. 1.8 is so quaint..

http://www.silentpcreview.com/article57-page5.html
His temps aren't outside the realm of possibilty.
The review of the 7000a gave it a 0.3°C/W at 5v. Multiply that by the 89 watts that Intel specs as the wattage, or the more "realistic" estimate of 100 watts, and you get a Delta of 26-30°

Doesn't the ARM case have a side duct over the CPU? If so then it's pulling in ambient air that's probably 20-22°. Add that to the 26-30° delta and you get a CPU temp of 46 to 52°. PLus there's also the fact that JohnMK's temps are not true full-load temps. A few hours of video coding is not the same amount of stress as CPUBurn would be
Last edited by Rusty075 on Mon Apr 26, 2004 8:42 am, edited 1 time in total.

Ralf Hutter
SPCR Reviewer
Posts: 8636
Joined: Sat Nov 23, 2002 6:33 am
Location: Sunny SoCal

Post by Ralf Hutter » Mon Apr 26, 2004 7:28 am

Rusty075 wrote:Plus there's also the fact that JohnMK's temps are not true full-load temps. A few hours of video coding is not the same amount of stress as CPUBurn would be
Ding-ding-ding!!! We have a winner here folks!!!

wumpus
Patron of SPCR
Posts: 946
Joined: Sat Sep 06, 2003 9:57 pm
Location: Berkeley, CA, USA
Contact:

Post by wumpus » Mon Apr 26, 2004 7:54 am

With a direct-to-outside-air side duct, that is possible. I hadn't considered that! But it is IMO the only way to achieve that temp @5v with P4-3.4.

Video encoding is in the same ballpark as Prime95 in my testing. However, I've stated my position on CPUBurn before: it's totally synthetic and unrealistic.

JohnMK
Posts: 159
Joined: Sun Aug 25, 2002 7:37 pm
Location: Seattle
Contact:

Post by JohnMK » Mon Apr 26, 2004 10:21 am

Ralf Hutter wrote:
Rusty075 wrote:Plus there's also the fact that JohnMK's temps are not true full-load temps. A few hours of video coding is not the same amount of stress as CPUBurn would be
Ding-ding-ding!!! We have a winner here folks!!!
What do you take me for? I know what I'm talking about when I report a statistic such as that. I also run two instances of SETI@home overnight to maximize register utilization on a Hyperthreading P4. The highest temp I have seen is 54C, overnight, running SETI@home. My room is approximately 72 degrees Fahrenheit. Video encoding with XviD & Avisynth utilizes two threads heavily and distributes work well across both virtual CPUs, similiarly well utilizing the registers of the P4 and getting my temps up just as high as SETI@home.

wumpus
Patron of SPCR
Posts: 946
Joined: Sat Sep 06, 2003 9:57 pm
Location: Berkeley, CA, USA
Contact:

Post by wumpus » Mon Apr 26, 2004 11:00 am

Well to be fair, I was the one questioning it, not those guys.

But if you have a fresh air duct to the CPU, such a temperature is indeed possible, so apologies for my skepticism.

I agree that video encoding is on par with Prime95, Seti, and Folding in terms of cpu temps. However, CPUBurn *will* produce higher temps than any other known app. That's part of the problem IMO.. it's a completely synthetic result that we can never duplicate with real workloads. Kinda silly if you ask me.

Rusty075
SPCR Reviewer
Posts: 4000
Joined: Sun Aug 11, 2002 3:26 pm
Location: Phoenix, AZ
Contact:

Post by Rusty075 » Mon Apr 26, 2004 12:02 pm

John, I wasn't meaning to demean your knowledge, really. My comment was just to point out that the temps from your loading would be different than what Mike got in his review, not that you didn't know what you were talking about. :lol:

CPUBurn is used for HSF testing specifically because it is unrealistic: its the worst-case scenario. There is nothing else that loads a CPU as completely as it does. It's like testing anything else; you don't do crash tests against a nerf wall, or test a fire extinguisher by blowing out candles on a birthday cake. :lol: It is also: simple, consistent (unlike F@H, or Seti, where the CPU isn't necessarily at 100% load 100% of the time), and tests only the CPU.

EDIT: Another reason to use whatever tool will get the CPU to its highest temp: Accuracy. We're limited (generally) to temperatures rounded to the nearest whole number. The greater the spread you can get between ambient and load, the more accurate the °C/W value you can get. We could, in theory, just test all the heatsinks at idle, but that would mask the small differences between heatsinks behind the rounding errors in the temp reporting.

JohnMK
Posts: 159
Joined: Sun Aug 25, 2002 7:37 pm
Location: Seattle
Contact:

Post by JohnMK » Mon Apr 26, 2004 1:14 pm

The difference will be very small. In any event I've shown that the Zalman CNPS 7000 AlCu cannot be ruled out as an effective cooling system, even at 5 volts, even on a P4 3.4GHz that's used for CPU-intensive tasks and gaming (ATI Radeon 9800 Pro All-In-Wonder).

Stevo@ARM
-- Vendor --
Posts: 162
Joined: Thu Dec 04, 2003 5:35 pm
Location: Rohnert Park, California, USA
Contact:

Post by Stevo@ARM » Mon Apr 26, 2004 4:34 pm

JohnMK wrote:The difference will be very small. In any event I've shown that the Zalman CNPS 7000 AlCu cannot be ruled out as an effective cooling system, even at 5 volts, even on a P4 3.4GHz that's used for CPU-intensive tasks and gaming (ATI Radeon 9800 Pro All-In-Wonder).
Very true JohnMK, and I am happy this discussion illustrates how well our ARM DIY StealthPC configuration performs in the real world with top of the line components. We spent years continually refining and upgrading this platform to handle these massive thermals and I am pleased it is working well for you.

wumpus I don't blame you for being a bit skeptical, from all the products we have evaluated over the years, our experience has been that most cases (and pre-built systems for that matter) from many vendors are usually over marketed and rarely live up to their hype. The difference is, we try very hard to deliver the genuine article. :wink:

So, just to cover all the bases, there is some real world variation from individual CPU to CPU (even at the same clock speeds) and other system components, and a reasonable margin of slack should be applied to account for those individual variations. JohnMK's figures are right on track from what we have seen in our test labs with our DIY Kit. To plan accordingly for a 3.2 to 3.4 CPU just factor 55C (+/- 3-4C) @ 22C ambient.

In a nutshell we have seen the exact same configuration running anywhere from 52C to 59C running the same tests at full load. All of Intel's cores are not created equal and some just get hotter than others, even at the same MHz. All that said, the DIY kit should do all of this for you including a high end video card and two Samsung SATA HDD at about 27-29dBA at full load depending on your individual unique circumstances.

Stevo

roaf85

Post by roaf85 » Mon Apr 26, 2004 5:19 pm

I still do not see how dell does it. Steve I am impressed with ARM systems. Do you test your computers against dell noise and temp wise? I would honest like to call down to ARM to get some more info. I seems like I have wrote a book on this forum and read over many threads and have yet to figure out an answer. Buy a Dell 8300 or build a computer on the Sonata platform. I am not going for complete silence just something that is not a jet engine like the computer I have now.

Post Reply