do amd's still run hotter than Intels?
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
do amd's still run hotter than Intels?
specifically, I'm looking to build a HTPC and due to price I would rather use AMD.
I know in the past AMD's required more power, thus put out more heat, but is that really the case today?
I'm looking at AMD 2400's and up to Barton 2800's (trying not to break the bank, but still get good speed) vs Intel P4 2.2-2.6
I'm looking for something that can be adequately cooled w/ a SLK-7 & panaflo - thus keeping the noise to a minmum.
Thanks
I know in the past AMD's required more power, thus put out more heat, but is that really the case today?
I'm looking at AMD 2400's and up to Barton 2800's (trying not to break the bank, but still get good speed) vs Intel P4 2.2-2.6
I'm looking for something that can be adequately cooled w/ a SLK-7 & panaflo - thus keeping the noise to a minmum.
Thanks
Re: do amd's still run hotter than Intels?
Xp2400 or a 2800 both put out a max of 68.3 watts of heat. While a P4 2.6 has a "Thermal Design Power" of 69W.
Intel stopped releasing the max wattage for their processors. Their reasoning was that since the P4 will just throttle down at high temps there was no point in designing heatsinks for the max theoricial heat output. My theory is that they knew that AMD's newer chips ran as cool or cooler as the Intels, they just didn't want to abandon that piece of deceptive marketing. it used to be that the "Thermal Design Power" was set as 75% of the max. If that's still close to accurate that gives the 2.6 a max of about 85 watts.
And when someone tries to tell you that the P4 is easier to cool because it has a "heat spreader" attached to it, ignore them. The cover is there to protect the chip during assembly, not to help with heat. Even Intel admits that the heat spreader actually impairs the cooling, they've said so in some of their developer documentation. It adds another layer of thermal interface, but the heat is still coming from an area about the same size an as AMD chip.
Ok, enough off topic ranting. Back to your processsor choice:
If I were you I'd take a look at an XP2600 Barton 333. You get the Barton core, with it's bigger cache, a 333Mhz FSB, and it costs about half what a P4 2.6 would. (On Pricewatch the 2600 shows at $95, and the P4 2.6 at $215)
An SLK-7 should work just fine.
Intel stopped releasing the max wattage for their processors. Their reasoning was that since the P4 will just throttle down at high temps there was no point in designing heatsinks for the max theoricial heat output. My theory is that they knew that AMD's newer chips ran as cool or cooler as the Intels, they just didn't want to abandon that piece of deceptive marketing. it used to be that the "Thermal Design Power" was set as 75% of the max. If that's still close to accurate that gives the 2.6 a max of about 85 watts.
And when someone tries to tell you that the P4 is easier to cool because it has a "heat spreader" attached to it, ignore them. The cover is there to protect the chip during assembly, not to help with heat. Even Intel admits that the heat spreader actually impairs the cooling, they've said so in some of their developer documentation. It adds another layer of thermal interface, but the heat is still coming from an area about the same size an as AMD chip.
Ok, enough off topic ranting. Back to your processsor choice:
If I were you I'd take a look at an XP2600 Barton 333. You get the Barton core, with it's bigger cache, a 333Mhz FSB, and it costs about half what a P4 2.6 would. (On Pricewatch the 2600 shows at $95, and the P4 2.6 at $215)
An SLK-7 should work just fine.
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Good summary Russ.
One more thing to add to your rant: The thermal diode that reads temps in the P4 is loaced in a corner of the core, literally as far from the hot part of the die as possible. While it may seem hard to believe, there are big gradients in temps in that tiny die, and the temp differential between the hottest spots and the temp diode location has been noted to be as high as 15C (see page two of the SPCR HS test methodology for a thermal map photo). All of which makes lots of people think that the location of that diode is no accident, Intel did it to make the P4 seem cooler than the AMD competition.
Their latest PSU design guide updates to increase 12V current requirements underlies the huge power needs of the P4s. See the middle of the Recommended PSU page.
One more thing to add to your rant: The thermal diode that reads temps in the P4 is loaced in a corner of the core, literally as far from the hot part of the die as possible. While it may seem hard to believe, there are big gradients in temps in that tiny die, and the temp differential between the hottest spots and the temp diode location has been noted to be as high as 15C (see page two of the SPCR HS test methodology for a thermal map photo). All of which makes lots of people think that the location of that diode is no accident, Intel did it to make the P4 seem cooler than the AMD competition.
Their latest PSU design guide updates to increase 12V current requirements underlies the huge power needs of the P4s. See the middle of the Recommended PSU page.
-
- Patron of SPCR
- Posts: 798
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Akron, OH (The Rubber Capital)
- Contact:
BTW, There is no "Barton 2600+ 333" Chip.Rusty075 wrote:If I were you I'd take a look at an XP2600 Barton 333. You get the Barton core, with it's bigger cache, a 333Mhz FSB, and it costs about half what a P4 2.6 would.
I assume you mean Barton 2500+
Barton only comes in 2500+, 2800+, 3000+ and 3200+.
There is a 333 Mhz XP2600+, but it lacks the extra L2 cache of the Barton. It's just a Tbred B@333.
I have my Barton 2500+ and I am very pleased with it. Cheap too.
I run it like a 2800+ with no extra heat.
I'm not that chinsy....1700...please : )
Actually I've got 2 machines one is still using the 1500xp...which I'll probably put a 2200 in since they're so cheap...everything else on the machine is ok. The barton will be for a new machine I'm about to build.
Quick question as it pertains to the barton 2500 and thoroughbred "b" chips:
aren't the new B chips and the bartons overclockable via the bios and or jumpers....meaning no pencil trick or mod to overclock like the earlier xp's?
Just wondering what your settings are to run the barton @ 2800
Actually I've got 2 machines one is still using the 1500xp...which I'll probably put a 2200 in since they're so cheap...everything else on the machine is ok. The barton will be for a new machine I'm about to build.
Quick question as it pertains to the barton 2500 and thoroughbred "b" chips:
aren't the new B chips and the bartons overclockable via the bios and or jumpers....meaning no pencil trick or mod to overclock like the earlier xp's?
Just wondering what your settings are to run the barton @ 2800
most nforce 2 boards at least can change the t-bred b multiplier with no physical changes to the processor. This can help you get higher FSB settings, but even without unlocking them you can overclock.
Intel procs on the other hand, have a locked multiplier; probably a bad move on their part since people would rather buy a 2.4C than a 3.0C because they'll get a much better overclock on the 2.4;
Intel procs on the other hand, have a locked multiplier; probably a bad move on their part since people would rather buy a 2.4C than a 3.0C because they'll get a much better overclock on the 2.4;
Not sure on your board choice (asuming you are nforcing2 for the int- audio on your HTPC) but as for the 2500+ overclocking word is they over clock good.
The T-breds overclock great, but with the I went with the "Barton" as the 2500+ is as cheap as can be right now. I got mine "retail" on New Egg for less than $100 when I got it, and they are cheaper now almost anywhere you shop.
The T-breds overclock great, but with the I went with the "Barton" as the 2500+ is as cheap as can be right now. I got mine "retail" on New Egg for less than $100 when I got it, and they are cheaper now almost anywhere you shop.
-
- Posts: 110
- Joined: Fri Apr 11, 2003 12:53 am
hmm I'm assuming you are digging that beer tab deeper wilst posting rus?Oops, I meant 2500+ Barton.
Hit the "6" when I meant to hit the "5". Why'd they have to put those keys so close together?
Speaking of which, guinness might make decent thermal goop but who's going to waste the really good stuff when as3 will do?
Cheers!
-Liq
-
- Patron of SPCR
- Posts: 700
- Joined: Thu Mar 13, 2003 2:38 pm
- Location: California, US
- Contact:
Re: do amd's still run hotter than Intels?
Just a warning--the SLK-7(00) will not accommodate an 80mm fan. The SK-7 will, and the SLK-800 will, but the SLK-700 won't. If you buy an SLK-700, be aware that the "L" stands for "L00ser!" and if you get the SK-7, you're a "ser!" which sounds like "sir", which I think is a much better thing to be. So be careful with your model numbers.Dr. Smith wrote:I'm looking for something that can be adequately cooled w/ a SLK-7 & panaflo - thus keeping the noise to a minmum.
-
- Posts: 105
- Joined: Mon Feb 10, 2003 4:04 pm
- Location: MN
-
- Patron of SPCR
- Posts: 700
- Joined: Thu Mar 13, 2003 2:38 pm
- Location: California, US
- Contact:
Gekkani wrote:101W?? Where say that?
Thermal Design Power is 82W
Rusty075 wrote: Intel stopped releasing the max wattage for their processors. Their reasoning was that since the P4 will just throttle down at high temps there was no point in designing heatsinks for the max theoricial heat output. My theory is that they knew that AMD's newer chips ran as cool or cooler as the Intels, they just didn't want to abandon that piece of deceptive marketing. it used to be that the "Thermal Design Power" was set as 75% of the max. If that's still close to accurate that gives the 2.6 a max of about 85 watts.
AMD runs to 90°C, Intel runs to 85°C (someone correct that if I don't recall correctly).
So you can run an AMD hotter. Which means you more easily get rid of heat, since delta-T counts - difference between ambient and heatsink temperature.
But! Intel shuts down a lot of the parts that isn't used, and draws less power. So if you're making a workstation and not a gaming station or CPU server that will run 100% or close, Intel will not use much energy at all.
You can do the same thing with an AMD, but it requires a chipset that supports it (VIA KTxxx does, AFAIK), and it's not supported and it can lead to crashes. Check out programs like VCool.
So you can run an AMD hotter. Which means you more easily get rid of heat, since delta-T counts - difference between ambient and heatsink temperature.
But! Intel shuts down a lot of the parts that isn't used, and draws less power. So if you're making a workstation and not a gaming station or CPU server that will run 100% or close, Intel will not use much energy at all.
You can do the same thing with an AMD, but it requires a chipset that supports it (VIA KTxxx does, AFAIK), and it's not supported and it can lead to crashes. Check out programs like VCool.
My gawd. FUD from MikeC?MikeC wrote: Their latest PSU design guide updates to increase 12V current requirements underlies the huge power needs of the P4s. See the middle of the Recommended PSU page.
Neither processor is particularly wonderful in the mips/watt rating used by lower power enthusiasts, nor are either processor particularly wonderful in the pr rating / model # / mhz / whatever metric you fud on per watt either. Both are close enough to being the same that it really does not matter. This has been pretty much true for a long time. It is not like the old days with the K6 vs pentium MMX.
The new power supply spec allows the motherboard to derive Vcore from +12v like server products have for years instead of +5v like desktop boards have for years. This reduces the size of the wiring from the power supply to the motherboard.
-
- Patron of SPCR
- Posts: 798
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Akron, OH (The Rubber Capital)
- Contact:
I don't think that qualifies as FUD.
For those not aware, FUD = Fear, Uncertainty, and Doubt. It has become generalized to refer to any kind of disinformation used as a competitive weapon.
I don't think there was any intention to mislead here, especially since from what I can tell, MikeC is an Intel fanboy
Lighten up.
For those not aware, FUD = Fear, Uncertainty, and Doubt. It has become generalized to refer to any kind of disinformation used as a competitive weapon.
I don't think there was any intention to mislead here, especially since from what I can tell, MikeC is an Intel fanboy
Lighten up.
I agree with the sentiment here that current AMD and Intel CPUs put out relatively comparable amounts of heat. However, I believe that the primary reason that Intel CPUs are perceived to run much cooler is because until relatively recently, the power saving HLT/STPGNT state (used when CPU is idle) of the Athlon CPUs was broken. On older Athlons with non-integer multipliers (i.e. 6.5x), there was a bug in the lower power mode that could cause a lockup.
Because of this problem, AMD apparently told the motherboard manufacturers to disable this feature. Recently, AMD has encouraged manufacturers to re-enable this feature, and some boards now enable this by default. I think Intel boards pretty much always had this feature working and enabled.
On my new nforce system (HLT/STPGNT enabled) with a XP1700 1.5v CPU clocked up to 2ghz (XP2400?), at idle, the low airflow heatsink barely gets warm. My setup behave just like an Intel P4 setup, cool at idle, hot when the CPU is working hard.
For those with older Athlon systems, or new systems that don't have this power saving state enabled, software utilities like CPUCool or VCool can be used to turn on this feature.
Because of this problem, AMD apparently told the motherboard manufacturers to disable this feature. Recently, AMD has encouraged manufacturers to re-enable this feature, and some boards now enable this by default. I think Intel boards pretty much always had this feature working and enabled.
On my new nforce system (HLT/STPGNT enabled) with a XP1700 1.5v CPU clocked up to 2ghz (XP2400?), at idle, the low airflow heatsink barely gets warm. My setup behave just like an Intel P4 setup, cool at idle, hot when the CPU is working hard.
For those with older Athlon systems, or new systems that don't have this power saving state enabled, software utilities like CPUCool or VCool can be used to turn on this feature.
AMD is starting to stretch their PR rating a bit for the top end models, though I don't think it's as extreme as you say, But, whether you call it out of whack or comparable depends on what application you're benchmarking. Athlons have always smoked P4s in business type applications, whereas P4s are vastly superior in SSE2 enabled applications (the Athlon is missing SSE2) and media encoding benchmarks.
Because the architectures of the P4 and Athlon are so different now, they have their individual strengths and weaknesses. To get the best bang for your buck, you need to figure out what kind of performance is important to you.
Here is what I think is a fairly balanced review of Athlon vs P4 performance:
http://www.tech-report.com/reviews/2003 ... dex.x?pg=1
Because the architectures of the P4 and Athlon are so different now, they have their individual strengths and weaknesses. To get the best bang for your buck, you need to figure out what kind of performance is important to you.
Here is what I think is a fairly balanced review of Athlon vs P4 performance:
http://www.tech-report.com/reviews/2003 ... dex.x?pg=1