Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: rsi_boy who wrote (162061)	3/13/2002 4:19:05 PM
From: Tony Viola	Respond to of 186894

This is progress? You have a 77% increase in clock speed and only a 40% increase in performance. Surely, if Intel had just build a 4 way xeon based on the P3-S core @ 1.4GHz it would blow away the pokey P4 xeon. 1. The PIII was maxed out in clock speed. In fact, Intel had a helluva time getting the 900 MHz out. 2. You don't scale performance % for % when increasing clock speed.

To: rsi_boy who wrote (162061)	3/13/2002 4:34:12 PM
From: Tenchusatsu	Respond to of 186894

rsi_boy, <You have a 77% increase in clock speed and only a 40% increase in performance.> A 40% performance improvement is nothing to scoff at. So what if Intel had to crank up the clock speed 77% to get there? The point is that they are there, and no one else is. <Surely, if Intel had just build a 4 way xeon based on the P3-S core @ 1.4GHz it would blow away the pokey P4 xeon.> P3-S is built on the 0.13u process. Xeon is built on the 0.18u process. But the 0.13u follow-on, code-named Gallatin, is coming late this year. It will feature double the L3 cache (from 1M to 2M) and higher clock speeds. I guarantee you that Gallatin would blow away any theoretical Xeon based off of the P3-S core. Tenchusatsu

To: rsi_boy who wrote (162061)	3/13/2002 4:39:38 PM
From: fingolfen	Respond to of 186894

This is progress? You have a 77% increase in clock speed and only a 40% increase in performance. Surely, if Intel had just build a 4 way xeon based on the P3-S core @ 1.4GHz it would blow away the pokey P4 xeon. Translation: eyes closed, ears covered I don't care that it beats the previous by 40%. I don't care that it beats Alpha. It's Intel and Intel sux because I'm an AMD fanboy and Hammer's the wave of the future, d00d! ... and I'm personally having really bad flashbacks to the old M.C. Hammer videos at this point...

To: rsi_boy who wrote (162061)	3/13/2002 4:57:04 PM
From: wanna_bmw	Read Replies (1) \| Respond to of 186894

Rsi, Re: ">Best 4-way P3 Xeon (900 MHz): 39158.09 >First 4-way Xeon (1.6 GHz): 55138.6 This is progress? You have a 77% increase in clock speed and only a 40% increase in performance. Surely, if Intel had just build a 4 way xeon based on the P3-S core @ 1.4GHz it would blow away the pokey P4 xeon." I understand where you are coming from, but there are several things you should keep in mind. 1) 40% is 40%, no matter how many more megahertz it took to get there. 2) For an I/O limited test like TPC, performance usually scales along with the number of processors, and with the memory bandwidth. It's hard to say how much came from the processor itself, but when in doubt, refer to 1). 3) The Pentium III Xeon had 2MB of outer level cache. The current Xeon MP has 1MB of outer level cache, so it has to work past this penalty. On the other hand, the Xeon MP has the benefits of much greater memory bandwidths, and Hyperthreading, so it's hard to say where the net benefit came from. But when in doubt, refer to 1). 4) The Pentium III-S has 512KB of second level cache, making it unfit for multiprocessor systems. If you were to use it in a 4-way or greater system, performance would drop off suddenly and unexpectedly. Besides, the CPU was designed and validated electrically with only two loads besides the Northbridge. Therefore, in order to design a 4-way system, you would need a complex dual FSB system, most likely with a large (4-8MB) off chip L3 cache to uphold performance. This would raise the cost of the system by a significant amount, which would still give the price/performance advantage to the Xeon MP system. Note that the IBM x440 system uses Xeon MP CPUs with the above configuration (despite the costs) to create an 8-way system. Apparently, they see an advantage to this over networking multiple 2-way FSB designs using the Pentium III-S. 5) Intel is being quite successful pushing the Netburst micro-architecture into the high end server space. Then again, they have very little competition. AMD wants to get Hammer in this space, but they have a long way to go, and you're not going to see competitive high-end systems from Rack Server or Boxx Technologies. wbmw