![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Anand re-examines Conroe benchmarks
- Thread starter jebo_4jc
- Start date
duby229 said:
I didn't understand what you were getting at before when you said "all they had to do was look at the tray". Now I understand. You're talking about temperature monitoring software.
duby229 said:
The thermal coefficient of a cooling device, along with inlet temperature(which is the temperature of the air drawn in by the heatsink) determines the temperature of the CPU. Since, cooling devices have different thermal coefficients the only thing "looking at a tray" is going to do is tell you what the embedded thermal sensor thinks the temperature of the CPU is. You *might* be able to infer such things as "the cooling is efficient", "the cpu temp is low at idle/under load", or "the temperature is about 66C". But, all other variables withheld, power consumption can't be measured this way. Only when the monitoring software also takes into account input voltage, output voltage, and amperage can a reasonable *guess* as to what power consumption is. These things have to be programmed into software for different processors/motherboards.
So no, even if Anand installed monitoring software, it would have told us nothing about power consumption. *maybe* it would give us the various temperatures - no point to that really.
However, there is an easy way to guestimate. Power consumption for a particular fabrication process = capacitance X frequency X voltage^2 X a fudge factor
Capacitance is a linear function of the number of transistors
The fudge factor can be neglected here.
Conroe looks to have a smaller die size than Presler. I should have less transistors. It's also lower frequency. It should be produced on a more mature process. It may run at lower voltage or the same but probably not higher than presler. Finally, there are power-saving enhancements to the cache that allow transistors not being used to switch-off.
Now, you can photoshop the CPU displays from IDF and extrapolate the die size. Along with the frequency, you can do a little algebra. If you can't, you're on permanent ignored until you learn algebra. I detest geeks who can't do basic math. Just kidding...maybe.
Next, the presler 3.0 has a thermal guideline of 95w. But, in reality it runs cool and therefore is expected to consume <95w. For brevity, use 95w.
As expected, the 2.66 Conroe should not be a big power hog.
RushFan
Gawd
- Joined
- Nov 3, 2005
- Messages
- 743
I've read all 17 pages and have 2 questions...
Where are the 64-bit application benchmarks comparing Conroe to FX-60?
Where are the multitasking benchmarks comparing Conroe to FX-60?
I need to know about these before I can make an informed decision about my next big upgrade.
Where are the 64-bit application benchmarks comparing Conroe to FX-60?
Where are the multitasking benchmarks comparing Conroe to FX-60?
I need to know about these before I can make an informed decision about my next big upgrade.
RushFan
Gawd
- Joined
- Nov 3, 2005
- Messages
- 743
savantu said:
Duh...the Anandtech article (see OP) CLEARLY was using an overclocked FX-60...why don't YOU do your homework.
...but you still didn't tell me anything about the 64-bit apps
And those tests were not multiTASKING (doing 2 different apps at once.)
MrWizard6600
Supreme [H]ardness
- Joined
- Jan 15, 2006
- Messages
- 5,791
savantu said:
yea right cuz thats the only diferance right, increase the clock speed and multiplier.
If you know the voltage of the chip, and the thermal dissapation of the heatsink, and the temp, of the chip, it is possibl to get a round about number for power, that should be within + or - 10%
But who wants to do that? Temp is close enough. The higher the temp, usually the higher the power. Of course that is not taking the thermal dissapation of the heatsink into account, but I dont really care too much. Most factory heatsinks actually do a pretty good job nowadays. They can be expected to work pretty well.
But who wants to do that? Temp is close enough. The higher the temp, usually the higher the power. Of course that is not taking the thermal dissapation of the heatsink into account, but I dont really care too much. Most factory heatsinks actually do a pretty good job nowadays. They can be expected to work pretty well.
(cf)Eclipse
Freelance Overclocker
- Joined
- Feb 18, 2003
- Messages
- 30,027
right, cause everyone knows you can oc your cpu to magically change socketssavantu said:
duby229 said:
So keep your "you have no clue" comments to yourself. Limted to two sockets, oh brother.
Sure they measure differently, doesn't change the end result. Oh wait, you swallowed the part about AMD saying their's is different because they have the Memory controller included? Dewd, I don't care. I go by the end results and they clearly showed Dothan and Yonah giving off less heat and using less power, deal with it? Conroe is supposed to have better Power Saving modes and might run as cool or cooler, sheesh!
Maybe showing temps couldn't be done because AMD biased folks like you would then say, "The FX-60 would have ran cooler if it weren't overclocked" or similar.
Intel also has better embedded processors hehehe!
Jerunk said:
I think it is Funny too folks call people Intel guys and even announce what they're running. My main system is an Asus A8N SLI and 3500+. Looking the biased AMD followers posts here, PLEASE add me to the Intel side, at lease they make sense.
Donnie27 said:
While it's a minor point... Intel has the memory controller in the northbridge, which produces heat seperately from the CPU, but still affects overall system temperature and power usage. AMD has the memory controller in the CPU, so it directly affects CPU temperature. I'm no expert on nano-tech power consumption, but I'd bet it counts for something... and may be a reason Intel chose to keep their memory controller out of the "core".
Different architechures are always going to be an apples to oranges comparison.
sabrewolf732
Supreme [H]ardness
- Joined
- Dec 6, 2004
- Messages
- 4,778
Donnie27 said:
erm, why would conroe run cooler than yonah? Conroe is a dekstop chip whereas yonah is a mobile chip. Don't understand why they would engineer conroe to run cooler than an already cool running chip
(cf)Eclipse
Freelance Overclocker
- Joined
- Feb 18, 2003
- Messages
- 30,027
i think he meant in comparison to K8..sabrewolf732 said:erm, why would conroe run cooler than yonah? Conroe is a dekstop chip whereas yonah is a mobile chip. Don't understand why they would engineer conroe to run cooler than an already cool running chip
sabrewolf732
Supreme [H]ardness
- Joined
- Dec 6, 2004
- Messages
- 4,778
(cf)Eclipse said:i think he meant in comparison to K8..
sabrewolf732
Supreme [H]ardness
- Joined
- Dec 6, 2004
- Messages
- 4,778
Donnie27 said:
it's funny how all the !!!!!!s always flame a company and then say "I run amd/intel so I'm unbiased"
I'm biased! I expect performance/cost radio to be as high as possible. I'm just not *platform* biased. Sure, I'm playing devil's advocate here sometimes, but that's only trying to swing the fanaticism back to realistic levels. There's a whole hell of a lot of assumptions going on in this thread... Some more far fetched than others.
Such as "Conroe will use less power than the mobile chips intel has been producing for 3 years because it's a new high-frequency desktop chip!" (misquoted, but you get my point.)
If Intel could have made a cool running desktop processor in the last 3 years, why didn't they? Oh yeah... cause they needed some !!!!!!s to pay for the R&D on the P4 before creating something that had tolerable power management levels.
Oh, and by the way... I don't give a crap about how many "speedstep" or "cache throttling" features a cpu has... I buy for gaming. So I want to know how much heat it produces running full throttle, for 2+ hour stretches. Any other power-saving features that cut performance are for laptop totin' sissies.
Such as "Conroe will use less power than the mobile chips intel has been producing for 3 years because it's a new high-frequency desktop chip!" (misquoted, but you get my point.)
If Intel could have made a cool running desktop processor in the last 3 years, why didn't they? Oh yeah... cause they needed some !!!!!!s to pay for the R&D on the P4 before creating something that had tolerable power management levels.
Oh, and by the way... I don't give a crap about how many "speedstep" or "cache throttling" features a cpu has... I buy for gaming. So I want to know how much heat it produces running full throttle, for 2+ hour stretches. Any other power-saving features that cut performance are for laptop totin' sissies.
I agree. Except for the laptop part
I'm not a gamer, but I got alot of heavy processes running, there are are times that I got 6 or more machines running full load for days. I want to know what the full load temp will be. The idle temp doesnt mean a whole lot to most people. Or a the very leat it doesnt mean mean a whole lot to IT guy's.
JetUsafMech
Gawd
- Joined
- Oct 19, 2001
- Messages
- 724
Logan321 said:
Yep, AMD has been burning up the innovation for quite some time now.
1:.................
2: Memory controller
3:.................
I have to agree with you, Intel was damn near chapter 11 before we !!!!!!s payed for the R&D on the P4. They are even back to the old tricks, whining before congress "Waaaaaa, monopolies, competition, we're innovative too, right? I mean, look at the built in memory controller, that's kick ass! What else? Well, ummm....we've paved the way for all the 64bit software that you can purchase at any fine retailer. Such as, you ask? Well..."
That was Intel whining for the whole world to see before congress, right?
JetUsafMech
Gawd
- Joined
- Oct 19, 2001
- Messages
- 724
Jason711 said:wow.. about the dumbest thing ive seen written. your top ten, for sure
Well don't leave me hanging, which part?
sabrewolf732 said:it's funny how all the !!!!!!s always flame a company and then say "I run amd/intel so I'm unbiased"
I'm not about to say I'm unbiased. The same biasedness that caused me to buy a 2.6C then caused me to buy an Athlon3500+. What this forum needs is a section for folks who use and like using both Intel and AMD systems. I wish you'd drop your personal thing you have against me
I will say you even seem a little more open than you use to be.I see something I like, because it fits a need or a like, then I become biased for that Part, no matter what it is. I don't care if the company sucks or not! When I built my last computer, I'd think if given a do over, I'd go with DFI instead of Asus and held out of a 7800GT instead of an X800XL, not change my 3500+ for Intel.
Logan321 said:
Conroe has more in common with the Dothan and Yonah than with P4. Timna made Intel gunshy. Dothan and Yonah made the whole memory controller issue moot. IMHO, Intel will more than like do an integrated memory controller on the Server Procs before their Desktop versions. Maybe Intel didn't want the controller there to save power draw and heat. But also costs as well.
I do fully agree with your last statement though. I'm honestly glad it is not apples to apples no choice = no fun! I like having both Intel and AMD systems because of that very reason.
Then get an XBOX LOL!
JetUsafMech
Gawd
- Joined
- Oct 19, 2001
- Messages
- 724
And 64bit has been around also. To give AMD credit, they brought both to the masses (desktop).Donnie27 said:
sabrewolf732
Supreme [H]ardness
- Joined
- Dec 6, 2004
- Messages
- 4,778
Donnie27 said:
Hehe same, I guess I seem to see donnie27 as author and I automatically think "intel !!!!!!" when I guess you're not really as !!!!!!ish as I think.
$BangforThe$
[H]ard|Gawd
- Joined
- Nov 5, 2005
- Messages
- 1,672
sabrewolf732 said:
Thats would seem to be the case donnie is a fence walker thats for sure. being neither hot nor cold.
I on the other hand am intel only. With me its not really about the most FPS its all around usage. Even thow I can afford to buy new every 6 months I don't . I buy for the long term . Intel to me with the northy's had a great cpu and platform to match. Thats not to say I haven't had setbacks but for the mostpart those were directly linked to MS.
I am getting pretty old so I want conroe to be perfect for me. I want this to be my last personnel computer. I also want a PC I can sell that I can point to a say this is as good as it gets period. The last generation had to many doudts for me from both Intel and AMD
Intel was a bit slower and heat was an issue . AMD have to many problems on the platform side of things.
Conroe seems to be the ans. I been waiting on . Performance Effiency and A great platform .
When I sell a man a $5000 gaminging PC I want to beable to look him straight in the eye's and say YA baby this is the real deal. I can't compete with DELL .
What I can do is Build a gamer that is = to or > than Voodoo / Falcon or alien ware. for less money . Price out the watercooled systems from these guys and add nothing but the best available hardware and see what they cost.
To me that counts for something.
sabrewolf732 said:erm, why would conroe run cooler than yonah? Conroe is a dekstop chip whereas yonah is a mobile chip. Don't understand why they would engineer conroe to run cooler than an already cool running chip
Please notice I said as cool or cooler not just cooler. Intel said they improved the process and Power Savings modes with Conroe's being much more advanced.
The whole new battleground is now Performance with Less Power and Heat or Performance Per Watt. AMD started it with the whole Cool & Quiet thingy.
So it was not just a matter of performance but Speed while sipping power and giving off as little heat as possible. So yes, Conroe will have every power and heat savings trick in the book. So I don't say what I said without good reason. So I agree with Duby in that I wanted to see the HD/PD results as well.
MrGuvernment
Fully [H]
- Joined
- Aug 3, 2004
- Messages
- 21,843
Jason711 said:educate me. i have no clue to what you are referencing.
Prior to Nforce4 - chipset problems - i am sure makers (dfi et cetera) comtributed but a AMD system never seemd to be put in all the parts, turn it on and off you go - seemed more like - turn it on, flash the bios, tweak the bios, then try this ram, or that ram, and eventually you can then install windows.
Nforce 4 was a god send - the amd systems i have @ work now are flawless! But the intel based ones have always been for me, assemble and go, done.
(cf)Eclipse
Freelance Overclocker
- Joined
- Feb 18, 2003
- Messages
- 30,027
i've never had a problem with any of the nf3 or nf2 boards i've ownedMrGuvernment said:
i think nvidia in general was a godsend to AMD's platform. how much work it takes to set up a system depends on the board you buy