anybody getting conroe?
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
anybody getting conroe?
Right now I have an X2 3800 cooled w/ an XP120
How do y'alll think heat output will compare to my X2 3800?
Thanks
Sean
How do y'alll think heat output will compare to my X2 3800?
Thanks
Sean
I was seriously considering it. I also have that cpu, I'm on the second motherboard I've tried that has sucked. Decided to try a third, arriving today. If it's not any better, I'm buying the first conroe ready board available and a cheaper p4 to last me until conroe's out.
The way it sounds, it'll be cooler, but not by a lot, like a quarter or a third less heat. But performance will be better. Kinda like comparing the x2 and the pentium d. My guess.
The way it sounds, it'll be cooler, but not by a lot, like a quarter or a third less heat. But performance will be better. Kinda like comparing the x2 and the pentium d. My guess.
-
- Moderator
- Posts: 7681
- Joined: Mon Dec 09, 2002 7:11 pm
- Location: Maynard, MA, Eaarth
- Contact:
Greetings,
There are many Athlon 64 X2 that have TDP less than 90watts.
http://forums.silentpcreview.com/viewtopic.php?t=30626
The other side of the coin is: is the 65watt Intel number truely equivalent? I.E. are they rating it in the same way as AMD does? Or, is it like the current Intel numbers, that is a fair bit lower than the actual maximums?
We won't know for sure how Conroe (or, as it will be known: "Core Duo 2") will compare until it becomes available, so there is no way that you can make statements about this.
There are many Athlon 64 X2 that have TDP less than 90watts.
http://forums.silentpcreview.com/viewtopic.php?t=30626
The other side of the coin is: is the 65watt Intel number truely equivalent? I.E. are they rating it in the same way as AMD does? Or, is it like the current Intel numbers, that is a fair bit lower than the actual maximums?
We won't know for sure how Conroe (or, as it will be known: "Core Duo 2") will compare until it becomes available, so there is no way that you can make statements about this.
-
- Posts: 84
- Joined: Thu Dec 01, 2005 7:39 pm
Or is could be like the mobile Intel processors, where the actual maximums are significantly lower than the TDP.NeilBlanchard wrote: The other side of the coin is: is the 65watt Intel number truely equivalent? I.E. are they rating it in the same way as AMD does? Or, is it like the current Intel numbers, that is a fair bit lower than the actual maximums?
It's much easier to compare TDP published by the manufacturers (Intel, AMD).
1) Power consumption of chips varies by their birthday, so different batch would have different overclock'ability and power consumption. and its further complicated by different minor steppings.
2) Most of CPU's power disspation/consumption are from switching power. With similar number of stages and similar clockspeeds, 65nm process will always cost a lot less power than 90nm process.
3) Process technology makes a huge difference too. AMD uses IBM's SOI process. Intel develops its own processes (it has 2 different 65nm processes with one high performance and one low power).
It's a safe bet to say Intel's upcoming CPUs will have higher performance and lower power consumption than all AMD's current offerings and next gen AM2 offerings.
1) Power consumption of chips varies by their birthday, so different batch would have different overclock'ability and power consumption. and its further complicated by different minor steppings.
2) Most of CPU's power disspation/consumption are from switching power. With similar number of stages and similar clockspeeds, 65nm process will always cost a lot less power than 90nm process.
3) Process technology makes a huge difference too. AMD uses IBM's SOI process. Intel develops its own processes (it has 2 different 65nm processes with one high performance and one low power).
It's a safe bet to say Intel's upcoming CPUs will have higher performance and lower power consumption than all AMD's current offerings and next gen AM2 offerings.
Which chips does this apply to? My pentium 4-M (not M) sometimes exceeds the TDP (30W) under full load.Or is could be like the mobile Intel processors, where the actual maximums are significantly lower than the TDP.
I thought with recent chips leakage power had started to approach 50% of total power? Or was that just Prescott?Most of CPU's power disspation/consumption are from switching power.
-
- Posts: 84
- Joined: Thu Dec 01, 2005 7:39 pm
-
- *Lifetime Patron*
- Posts: 1809
- Joined: Sat Apr 24, 2004 1:45 am
- Location: At Home
Conroe is going to easily beat a S939 X2 3800+ in the performance per watt stakes. When it comes to a 35W AM2 X2 3800+, it looks as if it’s going to be pretty close. This is at standard voltage; undervolting adds yet another unknown to the equation
If you buy an AM2 though, you have the chance of upgrading to a 65nm version in 6 months +, which might put AMD significantly back in front in the performance per watt stakes. Right now seems like a great time to speculate, but not necessarily to buy. The one thing I don’t like about the 35W AM2 part is that may turn out to be very hard to actually purchase, like Turion MTs in the UK or certain Opteron 1xx series chips. Saying that, it may well be that Conroe is also fairly hard to buy as a retail boxed chip for quite some time.
Yes, I do like to speculate
If you buy an AM2 though, you have the chance of upgrading to a 65nm version in 6 months +, which might put AMD significantly back in front in the performance per watt stakes. Right now seems like a great time to speculate, but not necessarily to buy. The one thing I don’t like about the 35W AM2 part is that may turn out to be very hard to actually purchase, like Turion MTs in the UK or certain Opteron 1xx series chips. Saying that, it may well be that Conroe is also fairly hard to buy as a retail boxed chip for quite some time.
Yes, I do like to speculate
It depends on process technology. SOI apperantly has less leakage but higher cost to design/manufacture. Intel's 90nm has some leakage problems. So did Broadcom's 90nm (manufactured in TSMC). I believe Intel's 65nm process(es) have less leakage problems compare to its 90nm.jaganath wrote:I thought with recent chips leakage power had started to approach 50% of total power? Or was that just Prescott?
-
- Posts: 220
- Joined: Wed Mar 15, 2006 10:16 am
- Location: U.S.A.
I think Intel is going to win the performance per watt crown at least until AMD releases the Revision G "Brisbane". The 35 watt 3800X2 might give AMD bragging rights but that is just one processor. Don't forget Merom either. Basically, all this speculation is giving me a headache lol.
EDIT: smilingcrow, I think I just paraphrased your post lol. I'm sorry bro, it's been a LONG day. Great minds think alike.
It appears Intel's Core 2 Duo will also be less expensive than AMD's AM2 Dual-Core. PCNation has the AM2 parts listed with their corresponding wattage and prices. I don't know how accurate it is though since there is a rumor of an AMD price drop on May 15. Check out the FX-62 with 125 watts @ 1.5volts . Also, see the change in wattage when AMD doubles the L2 cache. The 4200CUBOX is 65 watts whilst the 4400CSBOX is 89 watts. BTW, these launch on May 23.
sdat1333, I don't know if I'm getting Conroe or AM2 yet. I will wait until the smoke clears and then make a decision.
EDIT: smilingcrow, I think I just paraphrased your post lol. I'm sorry bro, it's been a LONG day. Great minds think alike.
It appears Intel's Core 2 Duo will also be less expensive than AMD's AM2 Dual-Core. PCNation has the AM2 parts listed with their corresponding wattage and prices. I don't know how accurate it is though since there is a rumor of an AMD price drop on May 15. Check out the FX-62 with 125 watts @ 1.5volts . Also, see the change in wattage when AMD doubles the L2 cache. The 4200CUBOX is 65 watts whilst the 4400CSBOX is 89 watts. BTW, these launch on May 23.
sdat1333, I don't know if I'm getting Conroe or AM2 yet. I will wait until the smoke clears and then make a decision.
From Hexus:
Intel's Conroe spanks AMD FX-62's botty - for real!!
Intel's Conroe spanks AMD FX-62's botty - for real!!
According to the benchmarks, the Conroe E6700 (2.67 Ghz) has a clear advantage over the FX-62 (2.8 GHz). Hexus compared the FX-62 to two Conroe's (E6600 @ 2.4 GHz and E6700 @ 2.67 GHz), an Athlon FX-60 and a Pentium EE 965. The Conroe E6700 beats the FX-62 in most of the tests. In ScienceMark, the FX-62 has a higher score, but otherwise, the Conroe seems to be the clear winner.What may also surprise you is that Intel already lists the unreleased Conroe on its web site, under the official branding of Core 2 Duo. The formal release is at least a month off but we, being an inquisitive and resourceful bunch of hacks, have managed to procure a couple of Conroe samples to test in our own (unbiased) lab.
They feature a dual-core architecture that's vaguely similar to Yonah's, with each core having access to a shared 4MB L2 cache. Pesky NDAs stop us from spilling the beans on just what makes the Conroe tick internally. What we can tell you is that the processors were clocked-in significantly below the 2.8GHz-rated Athlon 64 FX-62.
We can also confirm that the E6700 will ship with a 65W TDP (half that of Presler 965 and practically half FX-62's) and run on an Intel 975X motherboard with a Conroe-supporting BIOS and a couple of hardware changes - meaning that the 65nm CPUs will be released in LGA775 packaging. Sources indicate that the majority of 975X motherboards manufactured post-April 2006 have the necessary hardware modifications in place (a couple of resistors, supposedly). Intel is keen for Conroe to be a drop-in upgrade for its partners.
We've managed to run a complete set of benchmarks for both Conroe CPUs in our own lab environment guaranteed free of bias and we've compared their performance to the Presler Extreme Edition 965 and, uniquely, also to today's newcomer from AMD, the Athlon 64 FX-62 AM2.
Last edited by Slaugh on Sat May 27, 2006 7:44 pm, edited 1 time in total.
-
- Posts: 1608
- Joined: Tue Jan 04, 2005 4:02 pm
- Location: United States
Conroe is very tempting. I may wait a bit, though, and see what AMD 65nm offers. I have a feeling with SOI and their other technologies, they will become competitive again with Intel in performance per watt, but by then Intel will probably be close to rolling out 45nm, so who knows.
I just bought my dual-core a few months ago, so I'm going to try my best to hold out until quad-core shows up. But I'm kind of in the same boat as Sparkytfl, I have a motherboard that I'm not completely happy with. I could buy a new one, but if I'm going to do that, why not just go all out and get Conroe instead? From a performance per watt standpoint it should be a pretty big improvement.
I just bought my dual-core a few months ago, so I'm going to try my best to hold out until quad-core shows up. But I'm kind of in the same boat as Sparkytfl, I have a motherboard that I'm not completely happy with. I could buy a new one, but if I'm going to do that, why not just go all out and get Conroe instead? From a performance per watt standpoint it should be a pretty big improvement.
-
- *Lifetime Patron*
- Posts: 1809
- Joined: Sat Apr 24, 2004 1:45 am
- Location: At Home
I was very surprised to see that the 2.4 GHz Conroe also beat FX-62 by a comfortable margin overall. At this rate you could buy the cheapest Conroe, E6300 1.86 GHz $210, over-clock it to 1,333 FSB for 2.33 GHz and get performance similar to the $1,000 FX-62. That’s simply unheard of in my experience and with a TDP of roughly 50W for such an over-clock, it won’t be a beast to cool.
It should be possible to buy motherboards with official 1,333 FSB support as Conroe EE is touted as running at that speed, as do most of the Woodcrest Server chips. The 2MB cache of the slower Conroe chips versus the 4MB of the faster ones shouldn’t affect performance too much, except possibly when multi-tasking.
It should be possible to buy motherboards with official 1,333 FSB support as Conroe EE is touted as running at that speed, as do most of the Woodcrest Server chips. The 2MB cache of the slower Conroe chips versus the 4MB of the faster ones shouldn’t affect performance too much, except possibly when multi-tasking.
AFAIK the FX-62 is 2.8 GHz, not 2.6 as someone posted.
Also AFAIK Conroe has ~20% clock-per-clock advantage, so a 2.4 GHz Conroe should be just slightly faster than the 2.8 GHz FX-62.
I expect a price war around September, since both Intel and AMD will have plenty of CPUs on the market.
I doubt AMD will have the K8L available sooner than next summer, so Conroe will be quite attractive for a while.
As an AMD fan, Conroe might be my first Intel processor ever bought (in 14.5 years of PC ownership).
Also AFAIK Conroe has ~20% clock-per-clock advantage, so a 2.4 GHz Conroe should be just slightly faster than the 2.8 GHz FX-62.
I expect a price war around September, since both Intel and AMD will have plenty of CPUs on the market.
I doubt AMD will have the K8L available sooner than next summer, so Conroe will be quite attractive for a while.
As an AMD fan, Conroe might be my first Intel processor ever bought (in 14.5 years of PC ownership).
-
- Posts: 1386
- Joined: Fri Jun 18, 2004 6:53 pm
I'm seriously thinking about getting a C2.
My reasons:
1 - C2 performs much better than K8 at the same clock speed, but,
2 - C2 is still less expensive than K8 at the same clock speed. This doesn't mean that it will stay like this forever, AMD will lower their prices. But it's still a reason to wait a couple of months no matter which CPU you're getting.
3 - C2 runs cooler than K8 at the same clock speed, which is no surprise since it's made in 65 nm process. At least up to 3.4 GHz on air cooling with early engineering samples, the newer ones should go faster or at least make make sure that's a totally stable 3.4 GHz.
The biggest question is if I'll get a desktop or laptop PC...
About 35 W X2's, don't forget that the slowest C2 (1.6 GHz) will most likely run at a similar power usage since they have the same TDP = 65 W as the 2.67 W, and that's probably before heavy undervolting, unlike the low power X2 3800+, they are factory undervolted. Of course, the latter are hand picked. Still, I think the E4200 (1.6 GHz) is comparable to the X2 3800+, despite less L2 cache (2 MB, not 4 MB) and slower FSB. You could just pinmod it for just about any mobo though for higher speed. Only trouble with the E4200 is that it won't show up until Q4, but the price I've seen is $169.
X2 3800+ EESFF Vcore: 1.025V/1.075V
X2 3800+ EE Vcore: 1.20V/1.25V
X2 3800+ Vcore: 1.30V/1.35V
I think that SPCR's articles have shown that AMD's and Intel's TDP's are not far from the truth, just the fact that AMD uses individual TDP makes a bigger problem when comparing CPU's, even when comparing AMD to AMD. You just don't know until you've started it up.
My reasons:
1 - C2 performs much better than K8 at the same clock speed, but,
2 - C2 is still less expensive than K8 at the same clock speed. This doesn't mean that it will stay like this forever, AMD will lower their prices. But it's still a reason to wait a couple of months no matter which CPU you're getting.
3 - C2 runs cooler than K8 at the same clock speed, which is no surprise since it's made in 65 nm process. At least up to 3.4 GHz on air cooling with early engineering samples, the newer ones should go faster or at least make make sure that's a totally stable 3.4 GHz.
The biggest question is if I'll get a desktop or laptop PC...
About 35 W X2's, don't forget that the slowest C2 (1.6 GHz) will most likely run at a similar power usage since they have the same TDP = 65 W as the 2.67 W, and that's probably before heavy undervolting, unlike the low power X2 3800+, they are factory undervolted. Of course, the latter are hand picked. Still, I think the E4200 (1.6 GHz) is comparable to the X2 3800+, despite less L2 cache (2 MB, not 4 MB) and slower FSB. You could just pinmod it for just about any mobo though for higher speed. Only trouble with the E4200 is that it won't show up until Q4, but the price I've seen is $169.
X2 3800+ EESFF Vcore: 1.025V/1.075V
X2 3800+ EE Vcore: 1.20V/1.25V
X2 3800+ Vcore: 1.30V/1.35V
I think that SPCR's articles have shown that AMD's and Intel's TDP's are not far from the truth, just the fact that AMD uses individual TDP makes a bigger problem when comparing CPU's, even when comparing AMD to AMD. You just don't know until you've started it up.
-
- Posts: 125
- Joined: Sun May 14, 2006 5:30 am
- Location: Cheshire, England
-
- *Lifetime Patron*
- Posts: 1809
- Joined: Sat Apr 24, 2004 1:45 am
- Location: At Home
Interesting link Mats; I was wondering if there would be any under-volting headroom in the 35W X2 3800 and it does now seem very unlikely.Mats wrote: X2 3800+ EESFF Vcore: 1.025V/1.075V
This makes Core 2 Duo seem even more attractive, as an E6300 should be able to hit 35W with under-volting and give better performance; even after AMD drop their prices I can’t see them being competitive until 65nm.
The E6300 – E6700 Core 2 Duo (Conroe) are due to be released in July and hopefully will also actually be available then. Pricing starts at $210, they all have a FSB of 1066 and clock speed are 1.87 & 2.13 (2MB cache), 2.4 & 2.67 (4MB); whereas the E4200 has a FSB of 800, 2MB cache & 1.6 GHz. Mat’s was speculating purely on when the E4200 will be available as there doesn’t seem to be a firm date yet.whoatethepies wrote:Any solid news on when these are going to be released and the likely costs? Q4 is a little vauge.. I guess 'in time for Christmas'? Sure I read about an Intel chip that was due out in June/July...? Maybe I'm just going nuts
-
- Posts: 125
- Joined: Sun May 14, 2006 5:30 am
- Location: Cheshire, England
Cheers!smilingcrow wrote:The E6300 – E6700 Core 2 Duo (Conroe) are due to be released in July and hopefully will also actually be available then. Pricing starts at $210, they all have a FSB of 1066 and clock speed are 1.87 & 2.13 (2MB cache), 2.4 & 2.67 (4MB); whereas the E4200 has a FSB of 800, 2MB cache & 1.6 GHz. Mat’s was speculating purely on when the E4200 will be available as there doesn’t seem to be a firm date yet.
Personally I'd only look at the E4200 or E6600, all the others are too expensive in one way or another. What makes the E4200 so good is the low FSB, actually. It means that it got a multiplier = 8. The 6300 and 6400 are just harder to overclock with their low multipliers. I can raise the FSB myself, I don't have to pay for that. But I can't raise the multiplier. In the end I know that I don't need all that MHz anyway, at least not all the time.
The E6600 is the cheapest one with 4 MB L2 cache, and it remains to be seen if that's important or or not. I'll do what I always do, select the CPU type and the just pick the slowest model, unless MHz/$ tells me something else.
Now I don't know if this is up to date, especially not the prices.
Please note the different bus speeds. This is the reason why the only difference between a 4200 and a 6400 is a matter of BIOS settings, or even a pin mod if your BIOS don't have that function. The 6200 is just stupid, you need 500 MHz FSB to reach 3 GHz...
And then we have people who don't overclock, of course.
The 975X chipset can reach over 400 MHz.
I wonder how important L2 cache is for C2, does anybody know? All CPU's I've seen so far have 4 MB, even those who are labeled E6500 or lower, but that's just because they're ES.
The E6600 is the cheapest one with 4 MB L2 cache, and it remains to be seen if that's important or or not. I'll do what I always do, select the CPU type and the just pick the slowest model, unless MHz/$ tells me something else.
Code: Select all
Model Clock Cache FSB 200MHz 266MHz 333MHz 400MHz 500MHz Multi TDP Intro Price ($)
E4200 1,60GHz 2MB 800MHz 1600 2133 2667 3200 4000 8 65W Q4 169
E6200 1,60GHz 2MB 1066MHz 1200 1600 2000 2400 3000 6 65W Q4 179
E6300 1,86GHz 2MB 1066MHz 1400 1867 2333 2800 3500 7 65W Q3 209
E6400 2,13GHz 2MB 1066MHz 1600 2133 2667 3200 4000 8 65W Q3 239
E6500 2,40GHz 2MB 1066MHz 1800 2400 2997 3600 4500 9 65W Q4 269
E6600 2,40GHz 4MB 1066MHz 1800 2400 3000 3600 4500 9 65W Q3 309
E6700 2,67GHz 4MB 1066MHz 2000 2666 3333 4000 5000 10 65W Q3 529
E6800 2,93GHz 4MB 1066MHz 2200 2933 3667 4400 5500 11 ? Q4 749
E6900 3,20GHz 4MB 1066MHz 2400 3200 4000 4800 6000 12 ? Q4 969
E8000 3,33GHz 4MB 1333MHz 25? 95W Q4 1199
Please note the different bus speeds. This is the reason why the only difference between a 4200 and a 6400 is a matter of BIOS settings, or even a pin mod if your BIOS don't have that function. The 6200 is just stupid, you need 500 MHz FSB to reach 3 GHz...
And then we have people who don't overclock, of course.
The 975X chipset can reach over 400 MHz.
I wonder how important L2 cache is for C2, does anybody know? All CPU's I've seen so far have 4 MB, even those who are labeled E6500 or lower, but that's just because they're ES.
-
- *Lifetime Patron*
- Posts: 1809
- Joined: Sat Apr 24, 2004 1:45 am
- Location: At Home
Nice post Mats; I’m glad you didn’t forget the 99% of people that don’t over-clockMats wrote:The 6200 is just stupid, you need 500 MHz FSB to reach 3 GHz...
And then we have people who don't overclock, of course.
I think the E6200 is great for people who buy retail machines, as they’ll have a cheap low power CPU that doesn’t have a crippled FSB. I’m glad Intel is extending Core Duo 2 to lower price points.
As you stated, the E4200 does look a no-brainer for over-clocking on the cheap. I’ll probably go for an E6300, as I’ll hit my ideal power threshold before the FSB maxes out and I don’t want to wait for an unknown time period to save the $40.
I haven’t read any benchmarks, but typically for desktop applications the cache size advantage drops off pretty quickly. I can’t see it being an issue except maybe in a very heavy duty multitasking scenario.Mats wrote:I wonder how important L2 cache is for C2, does anybody know?
Yeah, the truth is that I don't care about them. "They" probably buy a Dell or similar anyway so "they" can't change FSB in BIOS. I didn't make that list for "them", I made it for those who are interested and know what I'm talking about.smilingcrow wrote:Nice post Mats; I’m glad you didn’t forget the 99% of people that don’t over-clockMats wrote:The 6200 is just stupid, you need 500 MHz FSB to reach 3 GHz...
And then we have people who don't overclock, of course.
AMD Opteron socket F will be launched on July 11.
Intel Core 2 Duo for desktop will be launched on July 23.
Looks like the prices i gave you are a bit wrong, The E6600 and E6700 are pretty much the same, but the E6300 and E6400 are priced lower at $183 and $224, respectively. This could mean that they won't make any slower models. Look at the new price for the PD 805, $93.
Intel Core 2 Duo for desktop will be launched on July 23.
Looks like the prices i gave you are a bit wrong, The E6600 and E6700 are pretty much the same, but the E6300 and E6400 are priced lower at $183 and $224, respectively. This could mean that they won't make any slower models. Look at the new price for the PD 805, $93.
-
- *Lifetime Patron*
- Posts: 1809
- Joined: Sat Apr 24, 2004 1:45 am
- Location: At Home
The pricing of the E6300 & E6400 is a pleasant surprise. I feel almost embarrassed for AMD at this point; the $183 E6300 over-clocked by 25% should match the $999 FX-62. I really hope they get their 65nm ramp together.
What’s a good chipset for over-clocking Core 2 Duo? I’m only looking for 1,333 FSB or there about.
What’s a good chipset for over-clocking Core 2 Duo? I’m only looking for 1,333 FSB or there about.
Please don't. You can't compare todays prices with a future price list. You won't get embarrassed, AMD will lower their prices when they have to. Until then they're still king of the hill (except for mobile CPU's). That's why I keep saying that you should wait 3 months before you buy a new CPU, no matter if it's coming from AMD or Intel.smilingcrow wrote:I feel almost embarrassed for AMD at this point;
Well that's the world of overclocking. But just to be fair you shouldn't compare stock to overclocked, it's already bad as it is with the FX-62 getting kicked by the E6600. Besides, we haven't seen the importance of L2 cache, and the E6300 only got 2 MB. Although I think it's pretty fast anyway. Dual core CPU's are getting lower prices, that's what happening. It's becoming mainstream. You saw that PD for $93, now that's a good price!smilingcrow wrote:the $183 E6300 over-clocked by 25% should match the $999 FX-62.
The only one right now is the high end 975X, but mobo maker Asrock seems to be able to make any chipset work, like 865.smilingcrow wrote:What’s a good chipset for over-clocking Core 2 Duo? I’m only looking for 1,333 FSB or there about.
However, the big mainstream chipset series is the 965 AFAIK. 333 MHz should be no problem with it.
Looking at the P-D price drops is anyone tempted to get a P-D 915 or 925 in mid July instead? They are supposed to drop down all the way below $150. Looks like it might make a good server box while you wait for Conroe MB/Chipsets etc to mature into early next year. Maybe its better to get a $200 conroe instead *shrug*.