fp 

nBLAST results

I just got done testing a Dual 1ghz G4 tower (because no xserves were available for me to test on) and discovered that nBLAST is significantly faster than a dual 1ghz intel box, however, it is only a touch faster than a dual 2.4ghz intel box.

which is great... right...

well not so great when you look at the prices of the two boxes. the dual 2.4ghz box is far cheaper than the xserve... when will apple learn...

it's not correct

to use a DP 1GHz tower to represent the Xserve. Besides the CPUs, the rest of Xserve's innards are superior to that of the tower.

Xserve vs. dual 2.4GHz

Is that dual 2.4GHz cheaper in a 1U rack config? Think BLAST farm. Think power and air conditioning requirements. Is it still cheaper?

Tower v XServe

It's completely wrong to assume the tower can stand in for the XServe in a test...the RAM in the XServe is DDR SDRAM, vs. the lowly PC133 in the Tower...I'm sure there are several other differences, but, also, keep in mind that the XServe isn't busy driving needless tasks that the consumer wants, freeing up more power to do server tasks.

blast

How many companies will be using blast in comparison to other programs out there? It's only a niche use in comparison to what other businesses need.

re: blast

True, Blast is niche, but Apache isn't.

Re: nBLAST Results

First, nBLAST is specifically optimized for cluster operation. Second, nBLAST is not Altivec-enabled. Had you used the Altivec-enabled version, the Power Mac would have blown the Intel box into the weeds instead or merely beating it by "a touch"...

Sheesh!

Re: nBLAST results

Look at the Xinet (xinet.com) results for an idea about Xserve vs. PMG4 performance.

Pretty...

I sure would like one of those servers... Even though I have no use for one. :

The guy who tested blast

I was the one who did the blast tests.

While the tower was not an Xserve, the Apple guy claimed it "would be just fine, maybe a touch slower".

I used the nBLAST that was altivec enhanced. it was only a touch faster. I sh*t you not. of course again, mhz for mhz apple blew the compition out of the water. I'm not disputing this. What I am going to say is that the 2.4ghz (1u) dual intel box did better when you factor in price. There just isnt a comparison at all. Apple charges too damn much, period. end of story. no really. Of course i could care less about independant ide channels when im talking BLAST, i mean sure i have databases that have to get into memory, but they can all fit on one drive.

To the guy talking about running other tasks, nothing else was running, i went in and manually killed jobs i didnt think were nessicary. This does not include aqua, it kept running, but i wasnt doing anything with it, noone touched the boxes while the proccesses ran.

I'd be interested in actually trying out an xserve, i really would, but apple was unable to get us any, even though we were talking about the possibility (if the price was right and the performance was there) of buying 20-30 of them pretty much right away. Apple just wasnt interested. Meanwhile our pc vendor, they drove the box to us...

I mean its simple really, the mac, sure it outperformed if you look at mhz, but if you factor in price, it was no contest, im writing up a purchase order of intel boxes right now.

THE MOST RECENT Intel roadmaps show that the firm is continuing to scale its Pentium 4 chip, with a 2.80GHz processor now slated for this quarter, quarter three.
This processor was shown in previous Intel roadmaps as a Q4 introduction.

Intel ascribes the shift to a "healthy" 130 nanometer process and good engineering. That means it's now highly likely that Intel will be able to introduce a 3GHz Pentium 4 for the fourth quarter, as the chip dates shuffle forward.

PC133?

The G4 tower with the dual 1GHZ processors has PC2700 DDR memory. The same that is in the Xserve.