Power Developer :: And the beat goes on ... ;-)

"Not that anybody on this thread gives a damn ... but I got an interesting email today from a developer who has a MP8641D dev kit. I have ars testbench 2.0 data now for MPC8641D, at 1.8 Ghz. On almost everything it is outrunning the dual (not dualcore) 1.8 970. On branchy scalar integer tests and on pseudo-random access tasks it is killing the 970. The bench doesn't directly measure latency, but the memory controller must be good and the bench numbers look "opteronish" in the way they scale with task footprint and from single-thread to dual. On latency-dependent dual-thread tasks significantly larger than Core1 caches ... it's beating core1 handily.

Single-thread FP is well down from 970 and moreso from Core1, no surprise.

It's a very potent altivec machine... beats everything except the dual-dual 970MPs running four threads. However most of the altivec in the testbench is well-optimized "

Author:	Karl [ Fri Sep 01, 2006 12:58 pm ]
Post subject:	And the beat goes on ... ;-)
Quote: "Not that anybody on this thread gives a damn ... but I got an interesting email today from a developer who has a MP8641D dev kit. I have ars testbench 2.0 data now for MPC8641D, at 1.8 Ghz. On almost everything it is outrunning the dual (not dualcore) 1.8 970. On branchy scalar integer tests and on pseudo-random access tasks it is killing the 970. The bench doesn't directly measure latency, but the memory controller must be good and the bench numbers look "opteronish" in the way they scale with task footprint and from single-thread to dual. On latency-dependent dual-thread tasks significantly larger than Core1 caches ... it's beating core1 handily. Single-thread FP is well down from 970 and moreso from Core1, no surprise. It's a very potent altivec machine... beats everything except the dual-dual 970MPs running four threads. However most of the altivec in the testbench is well-optimized " Source: ArsTechnica Forum Just for interest

Author:	bbrv [ Fri Sep 01, 2006 4:08 pm ]
Post subject:
The 8641D has promise. The board is big, but the chip looks good. We hope Freescale management starts to take an interest in altivec again! R&B

Author:	christian [ Sun Sep 03, 2006 3:19 pm ]
Post subject:
And what about a Pegasos with this state of the art processor ? Just in order to get a model between the "quad monster" and the actuel Peg ?

Author:	tarbos [ Sun Sep 03, 2006 9:38 pm ]
Post subject:
Imho the PWRficient holds more promise - if P.A. Semi can get it to market without unforeseen problems. Fortunately they both require similar board infrastructure with DDR2 and PCIe southbridge leading to a basic design/boardspec that could be modified easily to use the other processor respectively.

Author:	DJBase [ Mon Sep 04, 2006 7:28 am ]
Post subject:
A nice ÂµATX Board with such a CPU would be nice but like always - there is no desktop market for such a device. Everything is x86 related.

Power Developer https://powerdeveloper.org/forums/

And the beat goes on ... ;-) https://powerdeveloper.org/forums/viewtopic.php?f=3&t=764	Page 1 of 1

Author:	Karl [ Mon Sep 04, 2006 10:34 am ]
Post subject:
I would buy one, immediately. Also a PWRficient, but the 8641D is available, now

Author:	zylesea [ Mon Sep 04, 2006 4:56 pm ]
Post subject:
Quote: I would buy one, immediately. Also a PWRficient, but the 8641D is available, now Isn't it available in samples only yet? Nothing against Freescale, but their speed of pushing new products to market is not - cough - the fastest of the industry... Anyway, that first 8641D impression sounds promising. I hope Freescae roll it out soon and - even more important - roll it out with a competetive price tag.

Page 1 of 1	All times are UTC-06:00
Powered by phpBB® Forum Software © phpBB Group http://www.phpbb.com/