Are new G0 chips really too good OCers? - No!

[ziddey]

sure beats the hell out of 719a's

[Viss]

Around 545-550 booted ok and was good enought for some Spi1M ... have not clockgen`d further so far ...

If thats on air/water its more then enough! Thanks.

[Pete@X]

If thats on air/water its more then enough! Thanks.

All happend on water ... that reminds me -> I have to update my Sig

[hlonipha]

That might be the OCable difference between nvidia chipsets, and Intel chipsets

Wrong!!!!!!!, mine on a DFI 680i and after a few hours of non experience tweaking intel platforms, is at 3750mhz @1.39v on air 5 hours ORTHOS stable and counting as I left it running this morning, Memory is running at DDR556. We'll see when I get home from work.

[SLINROB]

How many fps do you gain going from 3.6Ghz to a 4Ghz CPU?!... 3, 5, 7 fps?!...

Not even that much in graphics heavy games such as Oblivion.

[SLINROB]

Its certainly not 'future proof' with the cheaper and cooler Penryn Quad only 6-8 months away.

Yes it most certainly is Their will always be something better around the corner so if you want to wait for all eternity be my guest.

Either way you look at it even though the penryn's will be better its not like the q6600 is a slacker and the most important benefit it offers are 4 cores (just like the penryn's) which future apps and games will be making use of very shortly.

[SLINROB]

Anyone know how much the Penryn Quad's are going to cost?

[SoulsCollective]

Yes it most certainly is Their will always be something better around the corner so if you want to wait for all eternity be my guest.

Either way you look at it even though the penryn's will be better its not like the q6600 is a slacker and the most important benefit it offers are 4 cores (just like the penryn's) which future apps and games will be making use of very shortly.

Well said.

[Tony]

regarding 680i Vs Intel chipset i hit just over 4.1 on the p35 boards and exactly the same on the DFI 680LT, the issue came when I tried to clock it with Striker Extreme...3.9GHZ was all I could get.

So its not so much chipset but board related, and how well they are made to clock high.

[OBR]

Lu(ky - eVGA is better than Striker ...

[SLINROB]

Guys how much do you think the 45nm quads will be when they first launch? I want a quad core cpu but I do not want to pay a fortune for one but on the same token I dont want to be missing out if I go with the Q6600.

I have a Tuniq Tower and a case with great airflow so as long as I can do at least 3.2ghz I will be happy. I have my $5 down at tankguys so I know eventually I will be able to go with the G0 stepping if thats what I decide to do.


This rig is solely for gaming and when games like crysis hit I am sure a quad core will be nice.

[Movieman]

Guys how much do you think the 45nm quads will be when they first launch? I want a quad core cpu but I do not want to pay a fortune for one but on the same token I dont want to be missing out if I go with the Q6600.

I have a Tuniq Tower and a case with great airflow so as long as I can do at least 3.2ghz I will be happy. I have my $5 down at tankguys so I know eventually I will be able to go with the G0 stepping.

I think you made the right choice. A Q6600 for app $300.00 shipped and a guaranteed G0 Step from Tank guys is the best deal going right now.
The 45nm stuff might do 10% clock for clock more but cost twice as much or close to that..
If I was in the market for a quad I'd do just what you did. It's the smart play.

[SLINROB]

I think you made the right choice. A Q6600 for app $300.00 shipped and a guaranteed G0 Step from Tank guys is the best deal going right now.
The 45nm stuff might do 10% clock for clock more but cost twice as much or close to that..
If I was in the market for a quad I'd do just what you did. It's the smart play.

So you think a 45nm quad will be much higher than $300 for a good while?

[Blauhung]

most likely the 45nm stuff will do the normal $1000 tray price for XE and work its way down much the same way conroe did when it hit. I would guess expect to pay $3-500 for a mainstream quad depending on what you go with

[SLINROB]

I would guess expect to pay $3-500 for a mainstream quad depending on what you go with

Yeah thats whats killing me, deciding to go q6600 right now or wait to see if I can get a mainstream 45nm quad for around $300.00


Will there even be any 45nm mainstream quads at first or just the expensive ones?

[Movieman]

So you think a 45nm quad will be much higher than $300 for a good while?

It's going to be Intels flagship and they will want a premium I'm sure.
If this was a change like from netburst to C2D then I'd say spend the extra but just a change from 65nm to 45nm with some small changes that could be offset by bumping up another couple hundred mhz I can't see spending $200-300 more..
I know, there is the "bragging rights" part of having a penryn but outside of the forum that and $2.00 will buy you coffee..
It's an individual decision but I like the G0 for $300.00. I think thats a steal for what your getting in computational power.
Run that up to just 3200 and you got one hell of a machine that can be cooled effectively on air..

[SLINROB]

It's going to be Intels flagship and they will want a premium I'm sure.
If this was a change like from netburst to C2D then I'd say spend the extra but just a change from 65nm to 45nm with some small changes that could be offset by bumping up another couple hundred mhz I can't see spending $200-300 more..
I know, there is the "bragging rights" part of having a penryn but outside of the forum that and $2.00 will buy you coffee..
It's an individual decision but I like the G0 for $300.00. I think thats a steal for what your getting in computational power.
Run that up to just 3200 and you got one hell of a machine that can be cooled effectively on air..

Sounds good and thanks for the advice I guess if it comes down to having to spend $500+ just on the mainstream 45nm quads then its a no brainer for me to save a couple hundred dollars and go with the G0 65nm Q6600.


Now come on Tankguys! hurry up and get those new steppings! LOL.

[Spanki]

It's going to be Intels flagship and they will want a premium I'm sure.
If this was a change like from netburst to C2D then I'd say spend the extra but just a change from 65nm to 45nm with some small changes that could be offset by bumping up another couple hundred mhz I can't see spending $200-300 more..
I know, there is the "bragging rights" part of having a penryn but outside of the forum that and $2.00 will buy you coffee..
It's an individual decision but I like the G0 for $300.00. I think thats a steal for what your getting in computational power.
Run that up to just 3200 and you got one hell of a machine that can be cooled effectively on air..

I'm not sure I disagree with your point of view or think you're wrong, but I'm not sure I agree with your characterization of the differences... Until the silicon hits the store shelves, noone will know, but dropping from 65nm to 45nm might well mean lower power consumption and/or lower voltage requirements, which could equate to better/cooler overclocks (they've already managed that with just a new stepping - the die-shrink may help even more).

I'd also not be so sure that they're going to charge a premium... they're not doing that for their current (and brand-spankin-new) FSB1333 'flagship' products.

Having said the above, it's really just a matter of time. If you're not in any hurry and can afford to wait til Oct./Nov., you'll have more info available and the current Q6600 may even be cheaper then (or at least bottomed out in price and older steppings syphoned out of the market). On the other hand, it's currently a bargain at the ~$300 price-point, so if your apps and usage patterns could benefit from a quad today (ie. encoding, 3d Apps, 1 or two games, or heavy multi-tasking, benchmarking/bragging-rights), then there's not much reason to wait.

The 'waiting' risk is.. you wait til Oct./Nov. at which point, you have enough info to make you want to 'wait til Feb 08', at which point you hear about something that makes you want to 'wait til April/May', etc. . At some point, you just have to pull that trigger.

[Movieman]

I'm not sure I disagree with your point of view or think you're wrong, but I'm not sure I agree with your characterization of the differences... Until the silicon hits the store shelves, noone will know, but dropping from 65nm to 45nm might well mean lower power consumption and/or lower voltage requirements, which could equate to better/cooler overclocks (they've already managed that with just a new stepping - the die-shrink may help even more).

I'd also not be so sure that they're going to charge a premium... they're not doing that for their current (and brand-spankin-new) FSB1333 'flagship' products.

Having said the above, it's really just a matter of time. If you're not in any hurry and can afford to wait til Oct./Nov., you'll have more info available and the current Q6600 may even be cheaper then (or at least bottomed out in price and older steppings syphoned out of the market). On the other hand, it's currently a bargain at the ~$300 price-point, so if your apps and usage patterns could benefit from a quad today (ie. encoding, 3d Apps, 1 or two games, or heavy multi-tasking, benchmarking/bragging-rights), then there's not much reason to wait.

The 'waiting' risk is.. you wait til Oct./Nov. at which point, you have enough info to make you want to 'wait til Feb 08', at which point you hear about something that makes you want to 'wait til April/May', etc. . At some point, you just have to pull that trigger.

I can't argue with any of your logic. You may well be right, I might be and it well could be a combo of both.
I agree, we'll see..but I still think the G0 Q6600 for $300.00 is a steal.

[ziddey]

I decided to give my l719a e6750 another go in my dfi p965-s. I tried for 6x425=2550 this time.

Set mode3/mode2 normal/normal 5-5-5-15-5-35 13-11-3-5
400bootup fsb
103pcie
1:1

1.5vcore, 1.325vtt, 1.35vnb, 2.125vdimm, 1.6vsb, 1.10vich

posting is sporadic, and windows locks up hard as soon as I start priming.

so i ram vtt to 1.5 and pump 1.7v to sb. now it's priming successfully.

Going to see if I can bring multi up to 7x and then 8x at these settings. Maybe the p965 dfi is limiting me and not the cpu itself (remaining optimistic )

[SLINROB]

I can't argue with any of your logic. You may well be right, I might be and it well could be a combo of both.
I agree, we'll see..but I still think the G0 Q6600 for $300.00 is a steal.

There is no doubt that the G0 quad at $300 is a still. Thank you guys for being so kind as to give me your 2 cents about the whole deal because this has been driving me nuts


After some thought I think I will just hold out as long as I can and first see how games such as crysis benefit from a quad core. By the time crysis comes out the 45nm's will be that much closer to release and the Q6600's should be that much cheaper.

Then again if the G0's turn out to be awesome overclockers and run fairly cool then it will be hard to say no when tankguys sends me a notification. LOL.


Either way I win

[ziddey]

Interesting. Used rmclock to switch multi up to 7x and then 8x and smallfft is still chugging along. I tried 1.5vtt by itself before but it didn't help. Guess the board needs 1.7vsb for some reason. Wonder how bad that is on longevity...

edit: dropped vtt down to 1.325v and it's still stable at 8x425=3400. tried clockgenning to 430 and got a hard lockup. maybe i'll get a new p35 board before giving up on this e6750.

I noticed thermal spec for these g0's is 72'c, and for b2's was 61.4'c. compensating by the 10.6'c makes coretemp reading identical to the reading from the motherboard's cpu readout.

That said, it looks like I'm loading around 50'c with poorly spread thermal compound (I was only quick testing so I didn't apply any compound, so whatever compound was what was stuck on the waterblock).

edit2: hmm. put all voltages to stock except vcore at 1.42v. wouldn't finish post. bumped vsb to 1.6v. vista crashes before the progress bar comes up. bump to 1.7v and vista almost gets to the gui but bsod's.

[gundersausage]

Only absolute top of the range insanely priced Penryns will be available on launch according to the Inq.. who knows when the affordable ones will kick in.. could be closer to mid year than start of the year before you can get one.. and thats without any delays.. Q6600 is here right now for chump change

[Spanki]

Only absolute top of the range insanely priced Penryns will be available on launch according to the Inq.. who knows when the affordable ones will kick in.. could be closer to mid year than start of the year before you can get one.. and thats without any delays.. Q6600 is here right now for chump change

True... my guess is March/April at the earliest for the affordable (sub-$500) 45nm quads - which is 8-9 months out from now. Even if they manage a Jan/Feb release, that's 6-7 months away.

Another consideration (one way or the other) is that those new 45nm parts will be the last of the socket 775 lineup. Sometime late '08 ('09?) Intel will introduce a new socket with an even newer cpu design.

So... you could upgrade to a Q6600 (or one of the E6x50 chips) now, and then 8mos. or a year from now grab one of the 45nm parts to ride out the rest of the S775 life-cycle (if it makes sense at the time - heck, maybe AMD will have pulled it's head out of it's ass (or it's ass out of the fire) by then).

[Blauhung]

I'd also not be so sure that they're going to charge a premium... they're not doing that for their current (and brand-spankin-new) FSB1333 'flagship' products.

Still going off a theory i have, but I really think that the G0 chips bin out much better from Intel's standpoint. So its much much cheaper to make them as you get many more mainstream binned die per wafer. I would venture a guess that the price points of the new 6*50 chips (even with the nice price cut when compared to the 6*00 stuff) have a very nice buffer against the cost of manufacture.

Can't really compare that to Penryn and it's still young process.

Edit:

True... my guess is March/April at the earliest for the affordable (sub-$500) 45nm quads - which is 8-9 months out from now. Even if they manage a Jan/Feb release, that's 6-7 months away.

Another consideration (one way or the other) is that those new 45nm parts will be the last of the socket 775 lineup. Sometime late '08 ('09?) Intel will introduce a new socket with an even newer cpu design.

So... you could upgrade to a Q6600 (or one of the E6x50 chips) now, and then 8mos. or a year from now grab one of the 45nm parts (if it makes sense at the time - heck, maybe AMD will have pulled it's head out of it's ass (or it's ass out of the fire) by then) to ride out the rest of the S775 life-cycle.

and I'm very not sure on this one, but i believe that mainstream Nahalem stuff that's still on a FSB could go without a socket change, but I only work in manufacturing so I don't hear about this stuff.