I815 or BX?

[Jkun]

I've decided to rip my trusty CUBX with 800 MHz PIII out of my capture box and make a HTPC around it.

To replace it, I will need a new BX or i815 board, with new processor.

The i815's 512MB limit isn't much of a problem. But the i815 boards are more expensive and don't have extra IDE / RAID controllers onboard like the CUBX-E. An extra IDE/Raid controller wil take up a PCI slot, as well as add to the price.

So what should I get? Is the i815 as trustworthy and stable as the BX?

Are there 1 GHz PIII's with 100MHz frontside bus? The only ones I could find were slot 1, not fcpga...

Will the new Tualatins work in both the BX boards and thew i815's?

J-kun

[Technoid]

If Tom's Hardware can be trusted, than the latest brand new I815 ep?? chipset mobo's will suport Tualatin (Easiest way to remember how to spell that name is "To late in").

1000Mhz or 1GHz FCPGA are available in 133Mhz.
Bus overklocking and capturing is a bad combination.

[Jkun]

<font face="Verdana, Arial, Helvetica" size="2">Originally posted by Technoid:
If Tom's Hardware can be trusted, than the latest brand new I815 ep?? chipset mobo's will suport Tualatin (Easiest way to remember how to spell that name is "To late in").

1000Mhz or 1GHz FCPGA are available in 133Mhz.
Bus overklocking and capturing is a bad combination.

</font>

Wasn't asking for overclocking but for use on a BX (100MHz bus only) Supposedly Tualatin would work on some BX boards even but lots of rumours but no facts.

J-kun

[Frank Marshall]

Aren't these very latest BX boards designed to operate at 133MHz FSB? I know the BX chipset wasn't intended to do this but according to Tom's Hardware, BX boards at 133MHz outperform all but the latest i820(?) chipset. So didn't Asus and couple other manufacturers simply build PC133 boards around the BX chipset with mobos like the CUBX? Even the name gives the intention away: CU = copper (coppermine) + BX. I could be wrong... it's happened before!

[Dr Mordrid]

I'll tell 'ya....Intel has got some real problems.

First of all they drop the ball when following up the BX. Then they produce the current P4 that itself will be out of date when a new version of it with a different pin format arrives this fall.

The uppance: those P4 boards now being sold to the unsuspecting public are not likely to be compatable with the new chips coming after the changeover.

Sheesh....

Dr. Mordrid


[This message has been edited by Dr Mordrid (edited 15 June 2001).]

[Jkun]

<font face="Verdana, Arial, Helvetica" size="2">Originally posted by Frank Marshall:
Aren't these very latest BX boards designed to operate at 133MHz FSB? I know the BX chipset wasn't intended to do this but according to Tom's Hardware, BX boards at 133MHz outperform all but the latest i820(?) chipset. So didn't Asus and couple other manufacturers simply build PC133 boards around the BX chipset with mobos like the CUBX? Even the name gives the intention away: CU = copper (coppermine) + BX. I could be wrong... it's happened before!</font>

Coppermines also exist in 100MHz format. The BX can be relatively safely OC'ed to 133, with a PCI divider of 4 in the new boards, the PCI bus runs at spec. However, the AGP does not, so basically I'd be overclocking my Marvel. Not a good idea.

Acc. to sandpile, there's only a 900MHz FCPGA CU, but even that one's not to be found. Shops I asked yesterday didn't have any 100MHz CU at all, and could no longer get those they claimed

J-kun

[This_Idiot]

<font face="Verdana, Arial, Helvetica" size="2">Originally posted by Dr Mordrid:
I'll tell 'ya....Intel has got some real problems.

First of all they drop the ball when following up the BX. Then they produce the current P4 that itself will be out of date when a new version of it with a different pin format arrives this fall.

The uppance: those P4 boards now being sold to the unsuspecting public are not likely to be compatable with the new chips coming after the changeover.

Sheesh....

Dr. Mordrid


[This message has been edited by Dr Mordrid (edited 15 June 2001).]</font>

Hang on Doc, you're sposed to be Pro-Intel. Sheet, if Doc can't recommend Intel OR AMD, where the hell are we all going to go ?

[Pace]

Crusoe? VIA Cyrix? Naah...

Maybe he's holding out for the NVIDIA CPU/GPU/POS combo

P.

[Dr Mordrid]

I'm not pro anything except for what will work for editing with the least muss & fuss. That's why I'm holding off on upgrading from BX/PIII.

Firstly, gaming performance is a total non-issue for me. I rarely game, and those games I like don't need any more performance than I already have.

While there are a lot of other *theoretically* faster options, well designed BX boards with a fast PIII are still the most stable, reliable video editing setups. A big part of this is that they have better PCI bus throughput mhz for mhz with far less bus noise than many supposedly faster rigs. Their only weaknesses are the lack of AGP4x, built-in 133mhz support for the AGP bus and ATA100, none of which are show-stoppers for editing.

The only thing I find remotely interesting right now is either a TBird with the AMD 760 or a dual TBird setup with the AMD 760MP. The single TBird looks interesting, but the 760MP is iffy right now since support for it is in the very early stages and its quirks, if any, are yet to be documented.

The current "replacements" for the BX are proving themselves to be painfully inept. The VIA issues aside there are major performance problems with the Ali MAGiK, SiS and many other "high performance" issuances.

The Taulatin's i815 B0 chipsets have the issue of integrated graphics, which I can't stand the idea of based on previous experience with such solutions. I was recently looking at the specs for the Asus TUSL2. This thing has integrated graphics, audio, Fasttrak100 Lite (no RAID0/1) and an ANR riser for networking. That's a lot to go wrong and if the BIOS disabling doesn't work the whole board's down the drain if something breaks.

Since there are several new chipsets on the horizon I find no real reason to hurry up the expense of updating 8 machines 'til they have a chance to prove (or DIS-prove) themselves.

As for the P4: it's still not performing up Intels' own expectations, so why "invest" in it?

Dr. Mordrid


[This message has been edited by Dr Mordrid (edited 17 June 2001).]

[dchip]

I have to agree with that. I'm still "plodding along" with a P-III 450 on a P3B-F and it works great. There is no reason to pursue the latest-greatest, especially when the performance and stability of the latest-greatest is questionable. Maybe I'll look at upgrading my processor to a faster
P-III before those get hard to find, but I'm staying with P-III and P3B-F.

[Dr Mordrid]

FYI,

my RT-2000 rig is running bios version 1006 and a PIII/850, which works great. I understand from others that the 1ghz Slot1 works with BIOS 1006 as well, if you can find one.

I've used both the Gigabyte 6R7 Pro/6R7+ and Asus Smart Slot 1 slotkeys with the Socket 370 CU-mines & P3B-F and both work.

Dr. Mordrid


[This message has been edited by Dr Mordrid (edited 17 June 2001).]

[Frank Marshall]

Yeah Jkun, overclocking the Marvel is probably not a good idea. I'm with ya on this, stability above all else... I wonder if the new 100MHz FSB Celeron IIs will eventually become the fastest processor for the BX PC100 platform? Aren't they up around 900MHz now and projected to break 1GHz before the year is out?

[Jkun]

Yes, the Celeron is up to 900 now, and will keep going up for at least 2 more steps before Intel will run into the same problems as the PIII and will have to change to .13 micron. If they then change to another busspeed too, remains to be seen.

The Tualatins look very good, and a few people have reported that dual Tualatins would be a real screamer, but Intel looks to block this because it makes their P4 looks poor. Pity.

Anyway, I asked in a separate thread if the Celly's would be just as good as a PIII for HuffYUV, depending on how much cache dependancy this codec has, but no reply. In general benchmark tests a PIII800 would still best a Celly 1 GHz...

J-kun