An un-busted Conroe-L (single-core Core Duo) got away!

All about them.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
Felger Carbon
Posts: 2049
Joined: Thu Dec 15, 2005 11:06 am
Location: Klamath Falls, OR

An un-busted Conroe-L (single-core Core Duo) got away!

Post by Felger Carbon » Fri Jun 29, 2007 7:23 am

The fabled Conroe-L (a single-core Core Duo) has surfaced. They tested it on "everyday" applications, and discovered it was about the same performance as dual-core Conroes. Which is why the final production form of the Conroe-L (to be known as the Celeron E400 series) will be seriously busted before the general public is allowed to touch it.

(An engineering sample Conroe-L somehow "got away".)

Here.

nd4spdbh
Posts: 88
Joined: Fri Dec 02, 2005 8:16 pm
Contact:

Post by nd4spdbh » Fri Jun 29, 2007 2:46 pm

you realize the newegg has had these for about 2 weeks now?

Celeron 420
http://www.newegg.com/Product/Product.a ... 6819116040

430
http://www.newegg.com/Product/Product.a ... 6819116039

440
http://www.newegg.com/Product/Product.a ... 6819116038

They all have a 35w tdp and clock very good. but in no means will they beat a a Core 2 duo.... as they are essientally 2 of these celerons on 1 die + some chache.

HueyCobra
Posts: 217
Joined: Tue Feb 14, 2006 12:16 am
Location: Australia

Post by HueyCobra » Fri Jun 29, 2007 5:56 pm

The Celeron 400s are also available in Australia.

X-bit labs previewed an engineering sample almost three months ago and posted very encouraging findings.
Ilya Gavrichenkov wrote:Almost in all tests Celeron 440 outperformed Celeron D 365 and in about half of all cases this advantage is more than 20% ... Note that besides the higher performance, the upcoming Celeron processor can also boast much lower heat dissipation and power consumption. The typical heat dissipation for the CPU on Conroe-L core will be 35W, while Celeron D processors on 65nm Cedar Mill core feature 65W TDP.
To my mind, there is no reason to choose a Celeron 300 over a Sempron. However, the competition between the Celeron 400 and Sempron LE (due around September) series looks like it will be very close. (Regarding the Celeron 400's 35W rating and the Sempron LE's 45W rating, note the difference in how Intel and AMD calculate/define TDP. Intel's ratings can be considered typical TDP, whereas AMD's ratings are maximum TDP.)

smilingcrow
*Lifetime Patron*
Posts: 1809
Joined: Sat Apr 24, 2004 1:45 am
Location: At Home

Post by smilingcrow » Sat Jun 30, 2007 6:55 am

HueyCobra wrote:the competition between the Celeron 400 and Sempron LE (due around September) series looks like it will be very close. (Regarding the Celeron 400's 35W rating and the Sempron LE's 45W rating, note the difference in how Intel and AMD calculate/define TDP. Intel's ratings can be considered typical TDP, whereas AMD's ratings are maximum TDP.)
This TDP issue gets mentioned a lot, but when I’ve looked at articles measuring real-world power consumption the TDP values of Intel and AMD seems to match pretty closely to power consumption for both brands. I know TDP isn’t the same as power consumption but the later is easier to determine and a useful indicator. Therefore I’ve concluded that it’s ‘fairly’ safe to directly compare AMD and Intel’s TDPs as a way of comparing their power consumption.

Bu the TDP by itself only tells a small part of the picture. You need to know the max clock speed that the TDP refers to and the IPC of the chips being compared. You also need to know the maximum temperature that the chip runs at before throttling itself.
e.g. if processor A has a higher TDP than processor B but also has a higher IPC and throttling temperature it could well be easier to cool quietly.

Well you might not need to know all these parameters, it depends what your use for the CPU is. :)

HueyCobra
Posts: 217
Joined: Tue Feb 14, 2006 12:16 am
Location: Australia

Post by HueyCobra » Sun Jul 01, 2007 7:05 am

smilingcrow wrote:This TDP issue gets mentioned a lot, but when I’ve looked at articles measuring real-world power consumption the TDP values of Intel and AMD seems to match pretty closely to power consumption for both brands. I know TDP isn’t the same as power consumption but the later is easier to determine and a useful indicator. Therefore I’ve concluded that it’s ‘fairly’ safe to directly compare AMD and Intel’s TDPs as a way of comparing their power consumption.
I'm inclined to think that there can be a meaningful difference, especially where AMD rates a group of processors the same regardless of the difference in clock speeds from bottom to top. Not only should the slowest processor differ (to a varying extent) from the fastest, but AMD's max TDP probably also includes some headroom and margin for error. Whereas Intel's typical TDP rating leaves the possibility for unpleasant surprises (examples can be found in Athlon 64 for Quiet Power), although it may be more or less accurate otherwise.

This shouldn't be much of an issue for budget CPUs at any rate, but I thought it was worth noting in case someone ruled out the Sempron LE for "running 10W hotter" than the Celeron E. I think the TDP of both CPUs will be similarly and satisfactorily low in typical use.
Last edited by HueyCobra on Sun Jul 01, 2007 1:06 pm, edited 1 time in total.

smilingcrow
*Lifetime Patron*
Posts: 1809
Joined: Sat Apr 24, 2004 1:45 am
Location: At Home

Post by smilingcrow » Sun Jul 01, 2007 7:42 am

HueyCobra wrote:I'm inclined to think that there can be a meaningful difference, especially where AMD rates a group of processors the same regardless of the difference in clock speeds from bottom to top. Not only should the slowest processor differ (to a varying extent) from the fastest, but AMD's max TDP probably also includes some headroom and margin for error. Whereas Intel's typical TDP rating leaves the possibility for unpleasant surprises (examples can be found in Athlon 64 for Quiet Power), although it may be more or less accurate otherwise.

This shouldn't be much of an issue for budget CPUs at any rate, but I thought it was worth noting in case someone ruled out the Sempron LE for "running 10W hotter" than the Celeron E. I think their TDP will be similarly and satisfactorily low in typical use.
I’m wondering how much performance can you get from AM2 and LGA775 CPUs using a Ninja+ and a Nexus 120mm before thermal throttling occurs. The test could be run with the Nexus set at a range between 5 and 12V or with another similar low noise fan.

Once you have the maximum CPU speeds you can then calculate the actual performance by using benchmark data published on hardware review sites.
This way you can bypass the TDP issue and look at how much actual performance you can get from each platform for a given cooling method, which also has a quantifiable noise level; use the SPCR data.
The test could be run with both a high end cooler and a low end cooler so you could see the performance difference you get at the same noise level threshold. The hard part about this is finding two fans of different sizes that have similar noise levels; I’m assuming this is needed as low end coolers will likely need a fan less than 120mm.

There are a few outstanding issues for me though:

1. Finding an LGA 775 motherboard that is 100% compatible with the Ninja+ (rev B). I would prefer to use the optional mounting bracket as I think the push/pull pin method is too flaky.
2. Which software monitors thermal throttling on the AMD platform? RMClock works okay for Intel but what is recommended for AMD and what is the thermal threshold for AM2 CPUs?

There’s probably no point in running this test at the moment at the high end because Intel should win easily. But using a low end quiet cooler might show some interesting result and once Phenom is released testing at the high end should be interesting.

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Sun Jul 01, 2007 8:58 am

smilingcrow wrote:
HueyCobra wrote:the competition between the Celeron 400 and Sempron LE (due around September) series looks like it will be very close. (Regarding the Celeron 400's 35W rating and the Sempron LE's 45W rating, note the difference in how Intel and AMD calculate/define TDP. Intel's ratings can be considered typical TDP, whereas AMD's ratings are maximum TDP.)
This TDP issue gets mentioned a lot, but when I’ve looked at articles measuring real-world power consumption the TDP values of Intel and AMD seems to match pretty closely to power consumption for both brands. I know TDP isn’t the same as power consumption but the later is easier to determine and a useful indicator. Therefore I’ve concluded that it’s ‘fairly’ safe to directly compare AMD and Intel’s TDPs as a way of comparing their power consumption.
That isn't a fair comparison at all. On the AMD chip is a dual channel memory controller. On the Intel chip there is no memory controller, the power usage for the memory controller is on the northbridge instead.

To make it a fair comparison you'd have to know exactly how much juice the memory controller on the K8 is drawing at any one time which is about impossible for a review site to get correct.

Doing whole system power comparisons end up comparing motherboards with widely varying features, manufacturing methods, part quality. Its great for real world comparisons but horrible for comparing processors accurately.

smilingcrow
*Lifetime Patron*
Posts: 1809
Joined: Sat Apr 24, 2004 1:45 am
Location: At Home

Post by smilingcrow » Sun Jul 01, 2007 10:10 am

dhanson865 wrote:That isn't a fair comparison at all. On the AMD chip is a dual channel memory controller. On the Intel chip there is no memory controller, the power usage for the memory controller is on the northbridge instead.
To make it a fair comparison you'd have to know exactly how much juice the memory controller on the K8 is drawing at any one time which is about impossible for a review site to get correct.
Doing whole system power comparisons end up comparing motherboards with widely varying features, manufacturing methods, part quality. Its great for real world comparisons but horrible for comparing processors accurately.
We are talking at cross purposes here.
I’m talking about using the TDP as a means of estimating how much performance you can get from a CPU when using a particular cooling method i.e. within a certain decibel threshold.
In this case the onboard memory controller is a hindrance as it increases the power density of the CPU compared to a system that uses the Northbridge.
I don’t really concern myself with the actual power consumption of the CPU but how much performance I can get from it for a given (silent) cooling method and still keep the CPU below the thermal throttling threshold.

For power consumption I’m only really interested at the platform level.

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Sun Jul 01, 2007 11:04 am

Sorry, the CPU cooling issues of a under 50W TDP are such a non issue to me I only think about power draw as related to overall cost of ownership and maybe the PSU fan ramping up which I have less control over. Huey's and your comments made me think that maybe one or both of you didn't know about the memory controller side of that equation.

Additionally I've only seen serious discussion of throttling on P4 class processors. I've never seen a K8 throttle, I know it has a shutdown temp but a throttle process on a desktop K8 is outside of my knowledge.

http://www.theinquirer.net/default.aspx?article=39894

Griffin can throttle itself up and down, and independently do the same for almost all IO mechanisms based on need. It can also do the same based on temperature. The older parts needed a separate thermal monitor, Griffin has them on die now. Additionally, the chip can monitor memory temperature and base memory speeds and usage on that. While this isn't anything new in the grand scheme of things, it is the first time it has been implemented on an AMD CPU.

http://www.tgdaily.com/content/view/32095/135
In 2008, the company will introduce the “Griffinâ€

smilingcrow
*Lifetime Patron*
Posts: 1809
Joined: Sat Apr 24, 2004 1:45 am
Location: At Home

Post by smilingcrow » Sun Jul 01, 2007 1:06 pm

dhanson865 wrote:Sorry, the CPU cooling issues of a under 50W TDP are such a non issue to me
One thing I’ve liked about AMD for years is that their stock coolers are very quiet when run at ~5-7V but still shift enough air to cool a CPU with ~35W. I have a Brisbane 1.9GHz undervolted that is cooled almost silently with the stock cooler.
dhanson865 wrote:Additionally I've only seen serious discussion of throttling on P4 class processors. I've never seen a K8 throttle, I know it has a shutdown temp but a throttle process on a desktop K8 is outside of my knowledge.
I’m a big fan of throttling as it allows you to not worry about the CPU temps. In other words it’s a backup in case your calibration was inaccurate or for whatever reason your CPU temps run wild.
dhanson865 wrote:Griffin can throttle itself up and down, and independently do the same for almost all IO mechanisms based on need. It can also do the same based on temperature. The older parts needed a separate thermal monitor, Griffin has them on die now. Additionally, the chip can monitor memory temperature and base memory speeds and usage on that. While this isn't anything new in the grand scheme of things, it is the first time it has been implemented on an AMD CPU.
http://www.tgdaily.com/content/view/32095/135
In 2008, the company will introduce the “Griffinâ€

HueyCobra
Posts: 217
Joined: Tue Feb 14, 2006 12:16 am
Location: Australia

Post by HueyCobra » Sun Jul 01, 2007 1:29 pm

dhanson865 wrote:Huey's and your comments made me think that maybe one or both of you didn't know about the memory controller side of that equation.
I'm aware that AMD processors have an on-die memory controller, but that's not relevant to my point in this discussion:
HueyCobra wrote:This shouldn't be much of an issue for budget CPUs at any rate, but I thought it was worth noting in case someone ruled out the Sempron LE for "running 10W hotter" than the Celeron E. I think the TDP of both CPUs will be similarly and satisfactorily low in typical use.

fri2219
Posts: 222
Joined: Mon Feb 05, 2007 4:14 pm
Location: Forkbomb, New South Wales

Data Point on Celeron-L 420

Post by fri2219 » Sun Jul 01, 2007 3:57 pm

I just dropped a 1.6 GHz Celeron L (SL9XP) into my file server, replacing a 3.2 GHz Celeron D (SL8HF).

BIOS reports 18C Idle/31C load temperatures, using a Zalman CNPS7700B-ALCU in an extensively modified P180. That's down from 32C/54C idle/load.

It's definitely a worthwhile upgrade from a thermal standpoint- the fact that dmesg reports a huge jump in BogoMIPS is a nice bonus.

If Noctua ever drops their prices, or Thermalright/Scythe dump their lousy LGA775 mounting solutions, I'll probably go fanless on my CPU heatsink.

HueyCobra
Posts: 217
Joined: Tue Feb 14, 2006 12:16 am
Location: Australia

Post by HueyCobra » Tue Jul 10, 2007 5:10 am

Legion Hardware has compared the Intel Celeron 440 vs. AMD Sempron 3600+:
Steven Walton wrote:On paper the Celeron 440 looks to be the superior processor and although the Sempron 3600+ dominated the memory bandwidth testing, it was the Celeron that claimed the processor performance testing ... However, once we started to get stuck into the 3D testing it became obvious that while the synthetic CPU tests did favour the Celeron 440, there was actually very little difference between the two processors.
Steven Walton wrote:Interestingly it was the Sempron that hit great heights with ease and despite its 62 watt thermal design, the AMD budget processor actually ran cooler than the Celeron when overclocked ... The Celeron heatsink ran at 46 degrees when the processor was under load, while the Sempron heatsink never got above 41 degrees while under full load operating at 2900MHz.
Steven Walton wrote:The Celeron 440 is a reasonably good processor at $75 US, but for a little more the dual-core Intel offerings are much better value ... The Sempron 3600+ on the other hand is simply amazing at just $50 US and with the kind of overclocking abilities that this processor has to offer, you can almost juice the full potential of a GeForce 8800 GTX. Therefore the Sempron 3600+ gets our pick, as this truly is a remarkable buy at just $50 US!

smilingcrow
*Lifetime Patron*
Posts: 1809
Joined: Sat Apr 24, 2004 1:45 am
Location: At Home

Post by smilingcrow » Thu Jul 12, 2007 1:22 am

HueyCobra wrote:Legion Hardware has compared the Intel Celeron 440 vs. AMD Sempron 3600+:
Interesting article and certainly surprising to me.
Steven Walton wrote:Interestingly it was the Sempron that hit great heights with ease and despite its 62 watt thermal design, the AMD budget processor actually ran cooler than the Celeron when overclocked ... The Celeron heatsink ran at 46 degrees when the processor was under load, while the Sempron heatsink never got above 41 degrees while under full load operating at 2900MHz.
Without knowing how the temperatures were measured I can’t see the relevance of this comparison. You have to consider the safe thermal limits for each CPU also and how these temperatures relate to those. Without knowing this data there is no context for this comparison.
My recent experience of stock AMD and Intel coolers has been that AMD has a much better product and that it runs very quiet at low RPMs. This would probably be enough for me to choose AMD over Intel if I was using a stock cooler.
Steven Walton wrote:The Celeron 440 is a reasonably good processor at $75 US, but for a little more the dual-core Intel offerings are much better value ... The Sempron 3600+ on the other hand is simply amazing at just $50 US and with the kind of overclocking abilities that this processor has to offer, you can almost juice the full potential of a GeForce 8800 GTX. Therefore the Sempron 3600+ gets our pick, as this truly is a remarkable buy at just $50 US!
It may be risky to purchase one of these Semprons without doing more research as it’s unclear how typical this over-clock is for this chip. The Intel seems to be a safe bet in over-clocking terms as do all Core 2 based chips.

A question for people with more experience of over-clocking AM2 systems than myself: Is a buss speed of 290 MHz fairly easy to achieve with budget AM2 boards?

smilingcrow
*Lifetime Patron*
Posts: 1809
Joined: Sat Apr 24, 2004 1:45 am
Location: At Home

Post by smilingcrow » Fri Jul 13, 2007 1:31 am

On second thoughts, I’d buy the AMD if I wanted to run at stock speed and the Intel as an over-clocker. I say that because you’ll have to disable CnQ if you overclock to 2.9GHz @ 1.5V and the power consumption will jump dramatically even at idle and with the amount of heat it will put out you’ll need to upgrade the cooler for it run silently.
Whereas you can typically hit 2.8GHz + with a Core 2 based CPU at stock voltage so you can still use EIST. I’d also upgrade the Intel cooler but it wouldn’t take much to cool it in comparison to the AMD @ 1.5V.

So for me it’s the same as with dual cores; AMD for value, low idle power and moderate performance and Intel if you need more oomph. I have an AM2 system because I don’t need top performance these days.

Post Reply