Oh, the chipset plays a huge role. The communication you speak of is (largely) regulated the NB in question. An exaggerated (because it has other problems) example is the 680i chipset. Even P965's are generally capable of running higher FSBs with QC chips than it can. Alternately, you can look at the RD600 chipset. It provides special tuning methods of the AGTL+ bus timings (only Core 2-compatible chipset that can do this as a tunable option as I understand), which have a -huge- impact on QC overclockability.
I still intend on getting the Q9450 or Q9550 when they're out, but probably won't be able to push it to the "limit" until P45. Yes, two dies will always limit what FSB you can hit, my argument is that given a chip technically able to hit the high FSB, the chipset plays a very substantial role. Probably even more important, however, is the motherboard itself...but given that P45 should give most makers their second stab at a design based on a chipset with native 1600QDR FSB support, I'm hopeful that we'll see the best QC OC results yet with it. Greg > -----Original Message----- > From: [EMAIL PROTECTED] [mailto:hardware- > [EMAIL PROTECTED] On Behalf Of James Boswell > Sent: Wednesday, February 20, 2008 6:16 AM > To: hardware@hardwaregroup.com > Subject: Re: [H] Need AMD XP Sempron CPU > > The issue with the twin-die arrangement on the current intel quadcores > is that it's not the northbridge that generates the FSB wall, it's the > blend of both cores and the need for everything to communicate > properly, I've had an E8400 on a 500Mhz FSB (P35 chipset) for a few > weeks now and it's very solid.... but the QC's seem to cap the FSB to > around 460-470 or so. > > It's feasible that no chipset will get around that, sadly :/ although > stranger things have happened. > > http://chryx.shacknet.nu/wolfdale4250.png > > > On 20 Feb 2008, at 03:35, Greg Sevart wrote: > > > I was considering moving to X48 in the coming weeks in anticipation > > of a > > 45nm quad-core upgrade (currently running a 65nm quad-core part at > > 3.2GHz on > > P35), but I think I'm going to hold off on the mobo move until P45. > > P45 > > sports ICH10R, which supposedly has 10GbE, and likely will do a > > better job > > overclocking the 45nm QC's. It's due out in June. > > > > The real problem is that the new Yorkfield series use a 1333QDR FSB, > > meaning > > that the multipliers are lower than the current Kentsfield. That > > means I'll > > need something likely capable of 475MHz+ FSB 24/7...that isn't P35 > > or X38, > > and likely not X48. P45 I think gives me the best chance at a high > > FSB to > > push to 4GHz. :) > > > > I wouldn't touch 780i with a 20 foot pole. > > > > Greg > > > >> -----Original Message----- > >> From: [EMAIL PROTECTED] [mailto:hardware- > >> [EMAIL PROTECTED] On Behalf Of j maccraw > >> Sent: Tuesday, February 19, 2008 7:33 PM > >> To: hardware@hardwaregroup.com > >> Subject: Re: [H] Need AMD XP Sempron CPU > >> > >> Thanks Duncan but it has to be 3000+ or better, 400Mhz > >> FSB part. =( > >> > >> LOL, you BiL is hard-up if he's hoping for charity in > >> the form of a free 1.2G > >> T-bird! Hell, I'm only doing replacement here because > >> the rest of the system > >> won't carry forward into X48/780i/DDR2/PCI-E land but > >> is more than usable for > >> HL2 & farcry the kids want to play. > >> > >> > >> DHSinclair wrote: > >>> j mccraw, > >>> I have the this for your consideration: > >>> -used- > >>> A1200AMS38 > >>> AXIA0108CPiW > >>> 95039650014 > >>> 1999 AMD > >>> > >>> Believe it should be an AMD Thunderbird 1.2GHz. > >>> If it will run in your m/b, it's yours....... I'll > >> deal with my B-i-L > >>> who seems to > >>> think he is going to get it someday out of family > >> gratis. > >>> Best, > >>> Duncan > >> > >> > >> > >> > >> > _______________________________________________________________________ > >> _____________ > >> Be a better friend, newshound, and > >> know-it-all with Yahoo! Mobile. Try it now. > >> http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ > > > > > >
