may i ask for some help?

[MDme]

Will be buying a quad core cpu for my bro and myself.

He wants a Q6600, G0 stepping: How do I find out what stepping the CPU is and if I buy from an etailer here in the US, how can I tell, I'm buying a G0 stepping.

I am waiting for a Penryn (45nm) CPU. Do you think it's advisable to wait? when is it coming out and what model would be a sweet spot? Q9450?

 

[lilbabimac]

I have the same question. I'm looking to buy a Q6600 also, but I'm wondering if it's worth waiting. For those of you who follow the trend of hardware prices, will the Q's prices drop enough with the release of Penryn to justify waiting til then? I need to get my rig setup asap but price is a big factor for me. Thanks.

 

[LOUISSSSS]

i know clubit.com has guaranteed g0 stepping for a pretty good price. get the OEM version + an arctic cooling freezer 7 pro with the $$ saved. stock cooler is a POS.

what i'm doing is get a ds3p (with 45nm and penryn support) & q6600 now. and use it until penryn really settles in. maybe after 2-3 price drops. (like i did with my AMD 4800+ and waiting for the Q6600 to be in the $250 price range.

p.s. newegg has also bee shipping g0 stepping lately

 

[myocardia]

Originally posted by: MDme
He wants a Q6600, G0 stepping: How do I find out what stepping the CPU is and if I buy from an etailer here in the US, how can I tell, I'm buying a G0 stepping.

You don't know what stepping you're getting, if you buy from most places. Only clubit and tankguys guarantee steppings. ClubIt is also slightly cheaper than newegg. I love the G0 Q6600 I got from ClubIt.

I am waiting for a Penryn (45nm) CPU. Do you think it's advisable to wait? when is it coming out and what model would be a sweet spot? Q9450?

It's only advisable to wait if you plan on doing mostly video encoding, or are going to try for 4 Ghz++ overclocks.

 

[toadeater]

Originally posted by: MDme
Will be buying a quad core cpu for my bro and myself.

He wants a Q6600, G0 stepping: How do I find out what stepping the CPU is and if I buy from an etailer here in the US, how can I tell, I'm buying a G0 stepping.

G0 stepping:

http://www.clubit.com/product_...l.cfm?itemno=CA1938452

I am waiting for a Penryn (45nm) CPU. Do you think it's advisable to wait? when is it coming out and what model would be a sweet spot? Q9450?

*If* you can wait, yeah, a Q9450 is better than the Q6600. It'll be more expensive too, because AMD dropped the ball with the Phenom.

 

[BonzaiDuck]

ClubIT is one place they advertise the processor and guarantee the stepping you want.

You can find others just by going to www.froogle.google.com and typing in "Q6600 G0" as all or part of your search keywords.

My own broader assessment of "waiting for Penryn" hinges on where you caught the "motherboard" wave, and how inclined you are to sell last year's board and buy a new one -- entailing a complete OS reinstall and file migration, for likely the chipsets would be different. More succinctly, it hinges on how compatible your current motherboard might be with the Penryn. For instance, there is a growing chance that the 680i boards will be compatible with the Wolfdale dual-cores, but not with the Yorkfield quad-cores. This almost seems to be a deliberate design-choice by INtel to spite nVidia, or nVidia's failure to get sufficient information from Intel in putting the 680i chipset together.

Logically, it seems that if a chipset like the 680i was compatible with Kentsfield, and if it's now certified compatible with Wolfdale, it would HAVE to be compatible with Yorkfield quad cores, but this isn't the story that has been circulating in recent weeks.

Even so, the Penryn processors -- say middle-of-the-line to second or third below the "flagship" model -- will be pricey for a while. In the meantime, you would probably see the price of the Q6700 come down, and the dual-core E6850 or even E6750 already look very promising. Just imagine: a dual-core that runs 1,333 FSB @ 3 Ghz that you should be able to over-clock easily to 3.6, with a cost between $170 to $300. And watch that change over the next several months, also -- for the better.

 

[BonzaiDuck]

But for sure -- anyone just migrating from AMD to Intel is in a pretty sweet situation. You can pick an INtel chipset from among several board manufacturers, and it would HAVE to be Yorkfield-capable.

So waiting for Penryn may depend on whether you want to spend Penryn prices, or are rather inclined to spend less on the most recent, 65nm incarnations of C2D and C2Q.

The G0 steppings go pretty far toward cooler, more over-clockable CPUs, even if Penryn is better.

 

[lilbabimac]

How about price wise? I'm not interested in the Q9450, just interested in how much it'll lower the Q6600 prices or even the E6850 prices.

*edit*
Thanks for the info Bonzai. I posted my reply above after you posted your replies.

 

[Denithor]

Since you are coming from single core, I would actually recommend something a bit different. Go with a motherboard with full Penryn quad support (IP35 or Pro, DS3L/R/etc) and then get an e21x0 now ($70-90). Build the rest of your system as desired (2/4GB DDR2, video card, etc) and then OC to 3GHz. This should provide enough power for you at the moment and you can upgrade to a Penryn quad mid/late next year once the prices are reasonable.

It looks like you are a gamer so also pick up a 3850/3870/8800GT and you have a fully loaded system ready to roll.

 

[bfdd]

Originally posted by: lilbabimac
How about price wise? I'm not interested in the Q9450, just interested in how much it'll lower the Q6600 prices or even the E6850 prices.

*edit*
Thanks for the info Bonzai. I posted my reply above after you posted your replies.

I doubt it'll lower the Q6600 price because it'll just be discontinued and the supply/demand will keep prices on it where they are.

 

[mwalt2]

How close will retail prices reflect the MSRP prices for the new cpus? Will it be like the current 8800GT prices (~$300 with an MSRP of $250) or will the prices posted so far on Penryns be close to what they'll actually sell for in January?

 

[MDme]

thanks for all the responses. I actually am debating between G0 Q6600 vs Q9450 from an OC standpoint (assuming the price is not too far apart). I am just wondering with the higher FSB and therefore lower multi how far an OC i can get with the Q9450...FSB already at 1333. It is 45nm though.

 

[Comdrpopnfresh]

I have a feeling that the q9450 will be released at the current q6600 price, and the q6600 will either stay there, or drop slightly (or be discontinued). It all depends on the improvements that penryn makes. They can in actuality sell penryn quads for less because they can fit more dies on a wafer. They will probably match the current price, as it seems to be a sweet-spot, and will actually have fairly large margins. By selling the q6600 for slightly less, they will slowly run out of those stocks, and will press more people towards the q9450 and get the higher margins. This can be seen with some of the pricing on newegg: The 6300 sells for the same or higher than the 6320, despite having more cache, and the 6550 sells for less than both- where it has more speed, the 4m cache, and the 1333mhz fsb! The same goes on as prices go up: 6750<6400, 6750<6600! For some reason, the xx50 variants, a refresh of the 6x00 line, sell for less, while having a faster fsb, and 260-340mhz more baseline speed. It penryn overclocks as well as everyone suspects, and there is a baseline improvement, I wouldn't be surprised to see higher speeds match current offerings, or same speeds being offered for less than current variants. This is how intel is able to keep their tick-tock motion going- amd on the other hand, is lowering their prices , on older offerings, and has high prices on new offerings (performance and other things aside)- it just isn't efficient for flooding the market with new offerings, which is a major stepping-stone to faster future offerings, and in the case of more efficient processes and die-shrinks, higher profit margins. The only way amd has done this is in killing 939. For this same reason, I suspect we'll see a refresh of the x48 platform with improvements (better ddr3 latencies and bandwidth usage), bug fixes (80mb/s sustained transfer rates), reduced current draw, and better overclocking around the time of the broad release of penryn. I also don't think we'll see penryn refreshes of the 21x0 line for a bit either- they haven't recieved bus upgrades or cache upgrades as the E4/E6/Q lines have, which only goes to herald how well Intel is doing in getting good yields.

 

[Comdrpopnfresh]

Originally posted by: MDme
thanks for all the responses. I actually am debating between G0 Q6600 vs Q9450 from an OC standpoint (assuming the price is not too far apart). I am just wondering with the higher FSB and therefore lower multi how far an OC i can get with the Q9450...FSB already at 1333. It is 45nm though.

I don't see why this is such an issue- my friend has a dfi lanparty 680i LT which "unlinks" the memory from the fsb- it always runs at 800mhz. Because of this, his q6600 is very easy to overclock. I mean, aren't people achieving the same basic 3.6ghz, fairly good max, on all the 21x0 lines? I don't think the multiplier is that important unless your ocing on the cheap. A good board can sustain high fsb speeds- and you're going to get heat increases either way...
And even so, in relative terms, a board that is not as good is not going to have the power regulation and stability to make the difference of a multiplier as profound anyhow.

 

[MDme]

Originally posted by: Comdrpopnfresh

Originally posted by: MDme
thanks for all the responses. I actually am debating between G0 Q6600 vs Q9450 from an OC standpoint (assuming the price is not too far apart). I am just wondering with the higher FSB and therefore lower multi how far an OC i can get with the Q9450...FSB already at 1333. It is 45nm though.

I don't see why this is such an issue- my friend has a dfi lanparty 680i LT which "unlinks" the memory from the fsb- it always runs at 800mhz. Because of this, his q6600 is very easy to overclock. I mean, aren't people achieving the same basic 3.6ghz, fairly good max, on all the 21x0 lines? I don't think the multiplier is that important unless your ocing on the cheap. A good board can sustain high fsb speeds- and you're going to get heat increases either way...
And even so, in relative terms, a board that is not as good is not going to have the power regulation and stability to make the difference of a multiplier as profound anyhow.

good points. but to clarify my point, the FSB speed of the CPU is what I am worried about. I mean to OC a significant amount, you would have to set the FSB of the CPU to insanely high levels like an FSB of 1600 or beyond. I wonder how the penryn generation will tolerate such FSB's.

going up to 1600 from 1333 = 20% OC
going up to 1333 from 1066 = 25% OC

Will the chipsets also tolerate that high of an FSB?

I may have some mistakes here since I haven't gone intel in a while.

 

[Comdrpopnfresh]

Well I believe some of the penryns will support a 1600mhz fsb- as many boards on the high-end are starting to offer it. But that will be a native feature- like boards running ddr2-1066 or 1200 even though the base fsb clock is still 200mhz or 333mhz. There will only really be a performance gain if the system setups can utilize the increased bandwidth- I don't think (please correct me if wrong) there is an inherent increase in actual speed- because all the benchmarks on 1333mhz buses vs the 1066 mhz buses only revealed increases in synthetic bench-tests, which mean relatively little besides the potential of a part.
If you want more speed, you have to increase the fsb from what the board offers, or the multiplier (in the case of unlocked processors). I can see where putting a 1333mhz fsb-enable processor on a 1600mhz board, because you know the board can handle the increase in fsb- much like putting 1066mhz ram on a board that only natively supports 800mhz; you know the ram won't be a limiting factor. That'll allow you to increase the fsb and the cpu will be the only limiting factor. But the overclocking potential of a 1600mhz fsb penryn will be strictly based on the architecture if placed on a 1600mhz enabled board.

 

[Denithor]

Um...I think you have missed something there, Comdr.

Overclocking results don't depend on the rated speeds for the board, just what the board and chip can actually handle.

For example, consider my rig. I have an e6400 (8x266 = 2.13GHz) on an IP35-E motherboard. I am currently running it at an FSB of 375 (8x375 = 3GHz) which in Intel's "quad pumped" nomenclature is equivalent to a bus speed of 1500mhz (4x375). This board is only technically "rated" to handle 1333mhz (333 fsb) but many people with the right RAM, chips, and cooling have hit or exceeded 500 fsb.

 

[Comdrpopnfresh]

I was just saying it's overclocking only if you go above stock speeds in all aspects, and in terms of actual speed that's done with the fsb and mulitplier (in the cpu). So having a higher initial fsb speed without an increase in core speed only increases bandwidth- you'd still have to increase the fsb further to speed up the processor. I mean- is a 3.0ghz core2 on a 1333mhz bus much faster than a 3.0ghz core2 on 1066? There is no inherent "speed" increase there, just bandwidth? Maybe I'm confused....