975 D0 Preview

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

SlowSpyder

Lifer
Jan 12, 2005
17,305
1,001
126
I am absolutely not jealous at all of that chip... I'm more than satisfied with my current set up. However, your water cooling set up on that case on the other hand.... yea.
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
20,839
3,174
126
Originally posted by: SlowSpyder
I am absolutely not jealous at all of that chip... I'm more than satisfied with my current set up. However, your water cooling set up on that case on the other hand.... yea.

i'll show off haruhi in cases and cooling once i get my final part.
 

chizow

Diamond Member
Jun 26, 2001
9,537
2
0
Just a quick update, MicroCenter restocked all of their Core i7 920 in stores, which sold out almost completely across the country. So if you were looking for a D0 and have a MicroCenter near you it may be worth taking a look. This was what I was waiting for and went in today to see if they had any D0.

They didn't. They had 45 something i7 processors at my location, one employee was pretty cool and helped me out, letting me sample 8-9 in the display case which were all SLBCH C0 steppings. They had another 4-5 cardboard shipping boxes with 5-6 CPUs in each which clearly had S-SPEC labeled on the big cardboard box to spare checking the entire box. All C0 steppings again. One thing I did notice however was that the pack date was only a few weeks ago, 03/18/09 with batch code 3843A.

I went ahead and got one as my other i7 is a 3842A with January pack date. The employee that helped me told me on the way up to the register that he and some of the other employees had already went through them all and didn't find any D0 SLBEJ. Now he tells me lol. Given my C0 was packed only 2 weeks ago, I'd say D0s are really just hitting the channel and if you hope to get one in the next month or so, a place like Tank Guys may be your best bet. The MC employee said he thought Friday at the earliest for their next shipment, but YMMV.

Anyways, my RAM came in from Newegg (DHL mailer FTL), so I'm going to go ahead and build with this C0. Worst case I just grab a D0 for $230 once MC gets them in if its really worth it, but I'll be happy with 3.8-4GHz with voltages at 1.35V or lower.
 

SlowSpyder

Lifer
Jan 12, 2005
17,305
1,001
126
Originally posted by: aigomorla
Originally posted by: SlowSpyder
I am absolutely not jealous at all of that chip... I'm more than satisfied with my current set up. However, your water cooling set up on that case on the other hand.... yea.

i'll show off haruhi in cases and cooling once i get my final part.

I would like to see what you can do. Pics when done, please. :) Every time I see a nice water cooling set up I end up at svc.com pricing things out. One day... but for now my Scythe Ninja copper will have to do.
 

Elfear

Diamond Member
May 30, 2004
7,096
640
126
Originally posted by: chizow
Just a quick update, MicroCenter restocked all of their Core i7 920 in stores, which sold out almost completely across the country. So if you were looking for a D0 and have a MicroCenter near you it may be worth taking a look. This was what I was waiting for and went in today to see if they had any D0.

They didn't. They had 45 something i7 processors at my location, one employee was pretty cool and helped me out, letting me sample 8-9 in the display case which were all SLBCH C0 steppings. They had another 4-5 cardboard shipping boxes with 5-6 CPUs in each which clearly had S-SPEC labeled on the big cardboard box to spare checking the entire box. All C0 steppings again. One thing I did notice however was that the pack date was only a few weeks ago, 03/18/09 with batch code 3843A.

I went ahead and got one as my other i7 is a 3842A with January pack date. The employee that helped me told me on the way up to the register that he and some of the other employees had already went through them all and didn't find any D0 SLBEJ. Now he tells me lol. Given my C0 was packed only 2 weeks ago, I'd say D0s are really just hitting the channel and if you hope to get one in the next month or so, a place like Tank Guys may be your best bet. The MC employee said he thought Friday at the earliest for their next shipment, but YMMV.

Anyways, my RAM came in from Newegg (DHL mailer FTL), so I'm going to go ahead and build with this C0. Worst case I just grab a D0 for $230 once MC gets them in if its really worth it, but I'll be happy with 3.8-4GHz with voltages at 1.35V or lower.

I'll be curious to see how your 3843A does. You may regret giving up the 3842A. A grabbed a couple i7s from MC and the best clocker was a 3842A (3.914Ghz@1.296V, HT on, air-cooling, ran Blend for 7+ hours). Post up some results when you get them.
 

chizow

Diamond Member
Jun 26, 2001
9,537
2
0
Originally posted by: Elfear
I'll be curious to see how your 3843A does. You may regret giving up the 3842A. A grabbed a couple i7s from MC and the best clocker was a 3842A (3.914Ghz@1.296V, HT on, air-cooling, ran Blend for 7+ hours). Post up some results when you get them.
So far it looks like my 3843A C0 is capable of similar clocks with HT on, although it takes more voltage @1.32V to get it stable. I don't think temps are the problem, as they don't break 64-67C on any core under OCCT with a through-bolted TRUE. Did you hit those speeds with 20 or 21 multiplier?

FWIW I think its VID is 1.21-1.23V. 1.21V was taken in BIOS, the 1.23V was the max VID registered in HW Monitor at all default settings, SpeedStep and Turbo Enabled while doing random Windows tasks and updates.

I haven't ruled out the BCLK being unstable on my board at higher settings, right now I'm testing 20x195 and its been stable so far for 30 mins in OCCT, HT testing enabled. If this turns out to be stable, I'll look at turning off HT and seeing if I can get 4GHz+ at similar voltages. I'll also consider running at 3.8GHz or so and seeing if I can get 1.28V or lower.

Overall it seems as if this chip is pretty typical for late-run C0/C1. I'd say Nehalem in general is much easier to overclock than Core 2, maybe similar to 45nm Penryn with regard to temps and max clockspeed headroom. The BIOS settings seem more streamlined and straightforward, with less need to tweak peripheral settings, although this could obviously be a limitation and shortcoming of my board. My initial impression is this Core i7 needs much less voltage to hit stable overclocks compared to my Q6600 and the reduction in heat is tangible.
 

Elfear

Diamond Member
May 30, 2004
7,096
640
126
Originally posted by: chizow
Originally posted by: Elfear
I'll be curious to see how your 3843A does. You may regret giving up the 3842A. A grabbed a couple i7s from MC and the best clocker was a 3842A (3.914Ghz@1.296V, HT on, air-cooling, ran Blend for 7+ hours). Post up some results when you get them.

So far it looks like my 3843A C0 is capable of similar clocks with HT on, although it takes more voltage @1.32V to get it stable. I don't think temps are the problem, as they don't break 64-67C on any core under OCCT with a through-bolted TRUE. Did you hit those speeds with 20 or 21 multiplier?

FWIW I think its VID is 1.21-1.23V. 1.21V was taken in BIOS, the 1.23V was the max VID registered in HW Monitor at all default settings, SpeedStep and Turbo Enabled while doing random Windows tasks and updates.

I haven't ruled out the BCLK being unstable on my board at higher settings, right now I'm testing 20x195 and its been stable so far for 30 mins in OCCT, HT testing enabled. If this turns out to be stable, I'll look at turning off HT and seeing if I can get 4GHz+ at similar voltages. I'll also consider running at 3.8GHz or so and seeing if I can get 1.28V or lower.

Overall it seems as if this chip is pretty typical for late-run C0/C1. I'd say Nehalem in general is much easier to overclock than Core 2, maybe similar to 45nm Penryn with regard to temps and max clockspeed headroom. The BIOS settings seem more streamlined and straightforward, with less need to tweak peripheral settings, although this could obviously be a limitation and shortcoming of my board. My initial impression is this Core i7 needs much less voltage to hit stable overclocks compared to my Q6600 and the reduction in heat is tangible.

I'm running mine at 19x206. I tried the 20x multi but the chip/board didn't seem to like it as well. I tried for 4GHz with HT on but it needed a lot of voltage to get semi-stable and temps were getting out of hand. Those are excellent temps by the way. Seems like the TRUE's do a good job of cooling the i7s. My Core Contact Freezer keeps things at mid to high 70s running Blend in the case.

I'll have to check my VID and see what it is.
 

chizow

Diamond Member
Jun 26, 2001
9,537
2
0
Gary Key just posted a D0 Preview. These are retail samples from Tank Guys, so no cherry-picking ES invovled. So far the results are impressive, 4.3GHz @ 1.32V on air, HT ON. Looks like D0 is a pretty substantial improvement in both clockspeed and voltages/temps.

I'll probably end up grabbing one when MC gets them consistently in a month or two, but this C0 isn't a slouch by any means. 3.9GHz right now with HT ON @ 1.32V so far, 21x186.
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
20,839
3,174
126
I TOLD YOU GUYS SO!!!

D0 = FTW!!!!!!!!!!!!!!!!!

They DONT run cooler tho. So if your getting it so it can run cooler, sorry, wait for the 32nm.

My sweet spot is 4.4ghz @ 1.35 vcore. :D

eVGA Classified coming in this week, so i hope to break 5ghz when that board arrives.

Originally posted by: chizow
Looks like D0 is a pretty substantial improvement in both clockspeed and voltages/temps.
.

Temps didnt change dude. It couldnt because the TDP is the same.

If your chip runs cooler its due to the lower voltage draw required on the chips, but u'll quickly find out with HT ON, + .05-.1 vcore @ 4ghz+ doesnt do much to a obnoxiously hot hardware, unless your on a low end air sink.
 

Rubycon

Madame President
Aug 10, 2005
17,768
485
126
Originally posted by: aigomorla
I TOLD YOU GUYS SO!!!

D0 = FTW!!!!!!!!!!!!!!!!!

My sweet spot is 4.4ghz @ 1.35 vcore. :D

eVGA Classified coming in this week, so i hope to break 5ghz when that board arrives.

I gave you a link to something that will certainly achieve 5G. ;)

I was going to get some D0 920's and now I probably will. My overclocking luck is not very good so I'm counting on them to be crappy. My 940 requires 1.525 volts to be stable at 4GHz with HT ON! That's pretty crappy eh? At that level it will throttle with TRUE cooling. Temps are nearly 30C cooler at 1.375V at 3.8 (HT ON). I can run at the same voltage at 4.0GHz with HT OFF! HT definitely changes things as far as stability goes when overclocking for max frequency.

Tell me what you think of the Classified. That's on my wish list as well. Someone's gonna have kittens when they see that bill because I always get TWO of everything in case something is DOA.
 

chizow

Diamond Member
Jun 26, 2001
9,537
2
0
Originally posted by: aigomorla
Temps didnt change dude. It couldnt because the TDP is the same.

If your chip runs cooler its due to the lower voltage draw required on the chips, but u'll quickly find out with HT ON, + .05-.1 vcore @ 4ghz+ doesnt do much to a obnoxiously hot hardware, unless your on a low end air sink.
Ya was referring to the reduced Vcore needed to hit 4GHz+, but I wasn't doubting D0 was a significant improvement, I was just trying to decide whether or not it was worth waiting a few days for a shot at one. Newegg using crappy DHL mailer made that decision easy for me until Saturday but I didn't plan on waiitng longer than that. Went ahead and installed the C0 I had and will just wait patiently until they get some D0 in stock. Of course I'll check again for D0 when I get around to returning the 2nd 920 C0 I have sitting here. :)
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
20,839
3,174
126
Originally posted by: chizow
Originally posted by: aigomorla
Temps didnt change dude. It couldnt because the TDP is the same.

If your chip runs cooler its due to the lower voltage draw required on the chips, but u'll quickly find out with HT ON, + .05-.1 vcore @ 4ghz+ doesnt do much to a obnoxiously hot hardware, unless your on a low end air sink.
Ya was referring to the reduced Vcore needed to hit 4GHz+, but I wasn't doubting D0 was a significant improvement, I was just trying to decide whether or not it was worth waiting a few days for a shot at one. Newegg using crappy DHL mailer made that decision easy for me until Saturday but I didn't plan on waiitng longer than that. Went ahead and installed the C0 I had and will just wait patiently until they get some D0 in stock. Of course I'll check again for D0 when I get around to returning the 2nd 920 C0 I have sitting here. :)

true, but unless your 4ghz+ OC is less then .05 -.1v you probably wont see a substaintial reduction in heat output.

Meaning simply, a 1.35vcore @ 4ghz will run near the same as a 1.3vcore @ 4ghz in aspect of heat. @ 1.25Vcore you'll see some reduction, but nothing that will determin stability over.

But a 1.2vcore 4ghz will run cooler significantly then a 1.35vcore.

Am i making sense?
 

chizow

Diamond Member
Jun 26, 2001
9,537
2
0
Originally posted by: aigomorla
Originally posted by: chizow
Originally posted by: aigomorla
Temps didnt change dude. It couldnt because the TDP is the same.

If your chip runs cooler its due to the lower voltage draw required on the chips, but u'll quickly find out with HT ON, + .05-.1 vcore @ 4ghz+ doesnt do much to a obnoxiously hot hardware, unless your on a low end air sink.
Ya was referring to the reduced Vcore needed to hit 4GHz+, but I wasn't doubting D0 was a significant improvement, I was just trying to decide whether or not it was worth waiting a few days for a shot at one. Newegg using crappy DHL mailer made that decision easy for me until Saturday but I didn't plan on waiitng longer than that. Went ahead and installed the C0 I had and will just wait patiently until they get some D0 in stock. Of course I'll check again for D0 when I get around to returning the 2nd 920 C0 I have sitting here. :)

true, but unless your 4ghz+ OC is less then .05 -.1v you probably wont see a substaintial reduction in heat output.

Meaning simply, a 1.35vcore @ 4ghz will run near the same as a 1.3vcore @ 4ghz in aspect of heat. @ 1.25Vcore you'll see some reduction, but nothing that will determin stability over.

But a 1.2vcore 4ghz will run cooler significantly then a 1.35vcore.

Am i making sense?
Yep yep makes total sense, temps don't really seem to go crazy until 4GHz+ AND 1.4V+, below that they're all similar under load. But the D0 may allow getting to 4GHz+ without needing those 1.4V+, which would be a nice improvement over C0. I still need to try turning off HT though and seeing max clocks with 1.4V or less. I know a lot of people turn off HT when reporting their clockspeed so it skews results a bit.
 

Rick James

Senior member
Feb 17, 2009
386
0
0
My C0 did 4ghz with everything on auto. Wish i had better cooling to see how high she'll go.
 

Gary Key

Senior member
Sep 23, 2005
866
0
0
Originally posted by: chizow
Gary Key just posted a D0 Preview. These are retail samples from Tank Guys, so no cherry-picking ES invovled. So far the results are impressive, 4.3GHz @ 1.32V on air, HT ON. Looks like D0 is a pretty substantial improvement in both clockspeed and voltages/temps.

I'll probably end up grabbing one when MC gets them consistently in a month or two, but this C0 isn't a slouch by any means. 3.9GHz right now with HT ON @ 1.32V so far, 21x186.

The preview is updated, I am running all tests currently at 4.41 on just slightly higher voltages but with memory still at 7-8-7-20 1T, just at 2100 now. I have the second processor going to Raja who will test it on the Classified board with his cascades and LN2.
 

Rubycon

Madame President
Aug 10, 2005
17,768
485
126
Originally posted by: palladium
Originally posted by: Rick James
My C0 did 4ghz with everything on auto. Wish i had better cooling to see how high she'll go.

What's the CPU-Z voltage like?

Ditto!

I was getting high clocks with AUTO and fired up CPU-Z and could not believe how much juice the board was giving the CPU! It varies with clock speed and when you're at 4GHz the VCORE was nearly 1.45V! :Q

That was Asus P6T6 WS Revolution with the latest BIOS. The original release BIOS supposedly overvolted even worse! :Q
 

Rick James

Senior member
Feb 17, 2009
386
0
0
Originally posted by: Rubycon
Originally posted by: palladium
Originally posted by: Rick James
My C0 did 4ghz with everything on auto. Wish i had better cooling to see how high she'll go.

What's the CPU-Z voltage like?

Ditto!

I was getting high clocks with AUTO and fired up CPU-Z and could not believe how much juice the board was giving the CPU! It varies with clock speed and when you're at 4GHz the VCORE was nearly 1.45V! :Q

That was Asus P6T6 WS Revolution with the latest BIOS. The original release BIOS supposedly overvolted even worse! :Q

Was getting 1.38v in cpuz with everything on auto.
 

The-Noid

Diamond Member
Nov 16, 2005
3,117
0
76
I think they definitely runa bit cooler. That was the reason I set mine up as well to keep it a bit cooler. Mine is for a media pc so it is not going to be overvolted at all so I am happy at 1.25v and 3.6 ghz with turbo enabled and a Zalman 9700 4 pin cooler.
 

chizow

Diamond Member
Jun 26, 2001
9,537
2
0
Originally posted by: Gary Key
The preview is updated, I am running all tests currently at 4.41 on just slightly higher voltages but with memory still at 7-8-7-20 1T, just at 2100 now. I have the second processor going to Raja who will test it on the Classified board with his cascades and LN2.
Very nice, I know in the preview you mentioned a custom D0 friendly BIOS for the Asrock board used. Do you think that would impact C0 overclocks at all? Also it'd be nice to see some results on other boards too, to see if there's any significant difference.

Originally posted by: Rubycon
Ditto!

I was getting high clocks with AUTO and fired up CPU-Z and could not believe how much juice the board was giving the CPU! It varies with clock speed and when you're at 4GHz the VCORE was nearly 1.45V! :Q

That was Asus P6T6 WS Revolution with the latest BIOS. The original release BIOS supposedly overvolted even worse! :Q
Yep and if you have Turbo on, it may overvolt even more. My board adds around 0.02V once the 21x Turbo multiplier kicks in.
 

Flipped Gazelle

Diamond Member
Sep 5, 2004
6,666
3
81
Originally posted by: chizow
Gary Key just posted a D0 Preview. These are retail samples from Tank Guys, so no cherry-picking ES invovled. So far the results are impressive, 4.3GHz @ 1.32V on air, HT ON. Looks like D0 is a pretty substantial improvement in both clockspeed and voltages/temps.

I'll probably end up grabbing one when MC gets them consistently in a month or two, but this C0 isn't a slouch by any means. 3.9GHz right now with HT ON @ 1.32V so far, 21x186.

That's just sick... :Q :D

 

Rubycon

Madame President
Aug 10, 2005
17,768
485
126
I prefer to keep turbo, speedstep, etc. disabled. Static frequency forever. :D
 

chizow

Diamond Member
Jun 26, 2001
9,537
2
0
Originally posted by: Rubycon
I prefer to keep turbo, speedstep, etc. disabled. Static frequency forever. :D
Ya I used to as well for the sake of stability, but from what I've read it doesn't really impact stability on the i7s. The C1E/Speedstep features are greatly improved on the i7s where it doesn't just clock the CPU down, it actually shuts down unused portions of the core. I have to admit, I've noticed the difference in ambient room temps from the power throttling features.

Plus it seems I actually need that 21x multiplier as my board (or RAM) tends to flake out around 200 Bclk. :)