Dual CPU MB's?

[Genom]

OK, fellow MURCERS, lend me your ears. After extensive thinking, I am kinda vetoing the idea originaly stated of getting a new faster CPU. Since I installed win2k a few days ago at home to start poking around, I am generally pretty impressed with it, having had no issues at all with a bunch of screwing around. Now, I have an old dual CPU MB with 2 P2-333's on it, so I know the pleasures of SMP (I use Lightwave a lot) and have been looking at getting an Abit Dual socket 370 MB, and getting 2 P3-550's to pop in there. Has anybody sued this MB? know anything about it? how are the PGA P3's performing? This solution would work great for me since i would get the performance boost I want, for considerably less than buying a 750Mhz P3. Dual CPU's and MB only cost 680 bucks, so it's a pretty compelling argument.

Any views on this would be greatly apreciated.


------------------
Asus P3B-F, P3-450~600, 256MB PC133 8ns SDRAM, G400MAX (yay!) at 160Mhz, Teac 58S CD-R, Toshiba SM-1002 DVD-ROM, Toshiba 40X SCSI CD-ROM, 8.4GB IDE primary drive, 9GB Micropolis Tomahawk UWSCSI, 9GB IBM U2WSCSI, 18GB U2WSCSI Cheetah, Diamond MX300, 3COM 10/100TX, 3COM Gaming Modem, Adaptec 2940U2W

And a partridge in a pear treeeeeeee.

[Guyver]

Interesting mispell...

Has anybody sued this MB?

If they used it and it sucked they might have!

To be honest, I haven't tried it, hopefully someone else here has... Someone has to have [bold]used[/bold] this type of mobo before...

Come on step up and admit it...

Guyv

[Thundrchez]

You're referring to the Abit BP6, eh? I thought that it was a dual celery motherboard only. Yes, it does use the same ppga cpus as the p3s now come in. The celerys do not natively do smp, and I think that they did a little bit of rewiring inside the motherboard to get the celery to do smp. This would probably make it incompatible with the p3s.

That is odd. Abit's webpage is showing it only as "Socket 370" processors, and not specifically celery. Can anybody else more difinitively state whether dual p3's work in an Abit BP6?

I've got a BP6 with dual celery. It is a bit of a nightmare wrt IRQs. I forget exactly what it is, but it is something similar to: AGP & PCI1 hardware share their irq. PCI2 has its own irq. PCI 3 is hardware shared with the integrated udma66 controller. PCI 4 & 5 are hardware shared with the USB. *ugh*

I distantly remember seeing a post or two regarding ACPI and the BP6. You may want to do a MURC search for those threads if you are thinking about w2k.

[Mark F]

Well acording to Abit the (current) BP6s will not work with the Flip-Chip (S370) P3s. However I saw an interesting bit of info on the HardOCP, about the newest BIOS containing code for the P3s. So maybe a future revision will work. Also some companies are looking to make a conveter from FC-PGA to PPGA (ie P3-type to celery-type S370).

Also current S370 P3s are supposedly not SMP capable
Mark F.

------------------
OH NO, my retractable cup holder swallowed a DVD...
and burped out a movie




[This message has been edited by Mark F (edited 23 February 2000).]

[jeepman]

I have a BP6 and I have been following the P3 issue since Friendtech announced their PL-S370 FC-PPGA to PGA convertor. Friendtech is using the BP6 as a proving ground, and they already have two P3 550E's running dual. The problem is they are developers' samples from Intel, and I still have not found any reliable info if the P3's publicly available are dual capable.

Intel's spec for the P3 500E and 550E say they are NOT dual capable, but Intel also stamped UNIPROCESSOR ONLY on the current Celerons and they still work dual. Only time will tell.

If you are not in a hurry, I would wait and see how the BP6 works out.

Besides the link to hardocp, check out http://ppga370.com and 2cpu.com and www.bp6.com for more info.

------------------
ABIT BP6, 192MB PC100 RAM, dual 366->>550MHz, 3DFXCool GlobalWin FEP32 'Lil Mofo' h/s fans. G400 DH, 3COM 905B NIC, Buslogic FlashPoint LW Ultra-Wide SCSI, SBLive PCI 1024. Segate Medalist PRO 9.1 GB UW SCSI 7200RPM, Maxtor 20.4GB ATA66 7200RPM 2MB Cache, 8GB tape b/u, IDE 32x CDROM, SCSI 4xwrite/8xread Panasonic CD writer, 4 more various 2GB SCSI drives.




[This message has been edited by jeepman (edited 24 February 2000).]

[Mark F]

How about this:

--------------------------------------------
Some more news on dually Coppermines on the BP6 from his Flyness:

Intel Claims to have hardware disabled SMP on their FC-PPGA CPU's. I personally think this may be an exaggeration on the part of Intel.
As we all know, i686 CPU's need pins #BR0, and #BR1, these two "Bus Request" pins allow two CPU's to speak with the bus, and each other to get a bus id, or CPU id. In the original white paper that Intel put out for the FC-PPGA at 500 and 550mhz, pin #BR1 was not defined as a pin. (Pin #BR1 is still missing from the Celeron white paper to this day.), however, about a month ago Intel put out the new white paper, Pentium® III Processor for the PGA370 Socket up to 750 MHz Datasheet. (found here: http://developer.Intel.com/design/pe...hts/245264.htm ). (notice this datasheets includes all FC-PPGA, even the 500mhz model.) Now we see that Intel added pin #BR1 at location X2. This is an issue on the BP6 since #BR1 is at AN15 on the Celeron. This would require an adapter to re-route this pin to it's new home at X2.

Intel further says: " The BR0# and BR1# (Bus Request) pins drive the BREQ[1:0]# signals in the system. The BREQ[1:0]# signals are interconnected in a rotating manner to individual processor pins. The table below gives the rotating interconnect between the processor and bus signals. During power-up configuration, the central agent must assert the BR0# bus signal. All symmetric agents sample their BR[1:0]# pins on active-to-inactive transition of RESET#. The pin on which the agent samples an active level determines its symmetric agent ID. All agents then configure their pins to match the appropriate bus signal protocol, as shown below. BR0# (I/O) and BR1# Signals Rotating Interconnect"

Now notice the reference to "symmetric agents ", that is Intel's way of not saying, additional CPU's, remember Symmetrical Multi-Processors, This is where Intel gets "Symmetric agents". They are referring to additional CPU's. Based on this information, I would guess that Intel didn't disable anything at all, they just moved pin #BR1, and didn't tell anyone. Remember that re-tooling these lines cost BIG bucks. In my opinion, it is unlikely that they would "hardware disable" the package, just to re-tool the line a few months later, but then again, they might. Only time will tell.
--------------------------------------------

Mark F.




------------------
OH NO, my retractable cup holder swallowed a DVD...
and burped out a movie

[birdy]

Jeepman,

If you run HD tach on your UDMA66 what CPU usage are you getting? I am getting about 75% where on my old board i was getting 2.5% UDMA33. Weird eh?

[birdy]

Forget that last question, there was a fault on the drive. It now works fine

Just made the leap to w2k,after a little trouble installing my diamond modem it seems to work fine. A bit quicker certainly on the net.

------------------
2 x Celeron 366@550 with huge HSFs. Abit BP6 MB. 128mb Winbond pc133, 128MB PC100 generic SDRAM, iwill 2930u scsi, diamond pro pci modem, SB Live!, pioneer dvd104s, hollywood plus dvd card, plextor 40x cd rom, ricoh 1420 cdr, ibm 13.5gb @ 7200rpm, G400 @ max settings. All wrapped up in a big case with loads of fans. Belinea 10 30 40 17" monitor