Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
DGagnon819, no problem. Here's the Inquirer article and the graphic of the improvements.
Paul
http://www.theinquirer.net/?article=10292
Delayed AMD64 compiler shows great promise
Is Intel about to feel Opteron's aftershocks?
By Mario Rodrigues: Wednesday 02 July 2003, 10:16
LAST MONDAY, the Portland Group, which has been working on compilers and other tools that have been optimized for AMD64, should have gone gold with PGI Workstation 5.0, which includes those optimizations, but instead announced a delay until on or before July 7.
For its unreleased product, the Portland Group is making performance improvements claims averaging 34% over PGI Workstation 4.1 on several industry standard benchmarks. Now that is a huge hike in performance from just compiler optimizations. The SPEC2000 floating point benchmark shows one Opteron result improving by almost 375%.
I can't predict the performance improvement gains that we'll see from recompiled applications, but they should be good. As Douglas Miles, director of the Portland Laboratory, has said :
"Porting our software, which represents well over a half million lines of code, from 32-bit X86 processor-based systems to 64-bit AMD Opteron processor-based systems has been very straightforward. We have also been extremely impressed with the performance of both legacy 32-bit applications and 64-bit applications re-compiled to take advantage of the new features of the AMD Opteron processor."
If we see large performance gains from recompiled apps, then the Portland Group would have delivered, and what I said in an earlier INQUIRER article would also have come to pass (Future Opteron enhancements, 3rd paragraph):
"Even though Opteron's performance has been nothing short of spectacular, I would characterize it as also being raw. Remember how SSE2 turned the P4 from an ugly duckling into a swan? Compiler optimizations that are being done by the Portland Group might change Opteron from being the best x86 server processor to one that looks beyond reach. The beta was released last April, so when it's ready for prime time this month, Intel had better taken out its Xeon and Itanium insurance cover. If the Portland Group delivers, Opteron's aftershocks may just push Santa Clara's enterprise group into the sea. OK! I'm being a little facetious, but you get the point I'm making."
AMD announced yesterday the immediate availability of its AMD Core Math Library (ACML) for its AMD64-based platforms, which will greatly aid the HPC community. This development tool helps increase code accuracy and speed of delivery for improved performance of 32-bit and 64-bit x86 applications and systems. With this announcement and the imminent AMD64 compiler debut, the AMD64 development tools have started to fall into place.
A raft of new benchmark results may also accompany the AMD64 compiler debut. If that is true, then this will be something to look forward to. If AMD's 2 GHz Opteron debuts as well, then that would be the icing on the cake. µ
http://www.pgroup.com/
http://www.pgroup.com/images/pg50vpg41.jpg
chipguy, re: "BTW, explain the business model that could possibly lead EA to develop a 64bit version of a
game with unique capabilities when there are hundreds of millions of powerful 32 bit PCs
out there and approximately zero AMD64 machines bought to play games?
Ever heard of plans for future growth? Gaming takes advantage of technical advances faster than any other segment of the industry.
Paul
This is encouraging, maybe astounding if it really pans out the way it sounds!
http://www.theinquirer.net/?article=10292
Paul
EP, I'd agree - I think Hector's done all right. Now, if things don't look considerably better a year from now...
A corporation the size of AMD(no, I'm not saying they're huge) turns more like an aircraft carrier than a destroyer. There are too many foundations that have been laid that can't just be abandoned without validated replacements. If the proper guidance has been instituted, the results will show up before too long.
Paul
yb, the review seems not to have been properly proofread, it has several obvious errors in it.
Paul
re: A64 notebook
"The game ran for several hours straight during the course of AMD's presentation and not only maintained a smooth frame rate, but did so without crashing or displaying any kind of visual glitching. For a beta platform running a beta-OS and using early drivers, this is a great feat and demonstrates AMD's commitment to making sure the entire platform, not just the CPU, is ready to roll come September.
Good news! That 16" LCD must look nice!
Paul
http://www.sudhian.com/showdocs.cfm?aid=405&pid=1445
drjohn, don't put too much stock into his review, he seems to confuse the data for the two chips a few times. I'd wait for a more capable reviewer before I assumed too much.
Paul
burn2learn - ever heard of Fab30? Why is it so hard to conceive of AMD picking the useful, appropriate techniques they can garner from IBM that will pan out for AMD's purposes? Last I checked, IBM was not AMD's sole output producer. For that matter, have they even planned to actually produce anything for AMD yet?
Paul
thanks, Semi!
UpNDown, do we know for sure that A64 is pin incompatible with socket 940? Obviously the inverse is incompatible, but has anyone analyzed the actual pinout to verify that A64's pins aren't all supported properly in socket 940?
Semi, so the cost advantage more than makes up for the lower # of dice per wafer?
sgolds, from the part #, the voltage is 1.6, so it's not likely a mobile part.
Paul
EP, good point about forecasting demand for the different dies. I'd forgotten to consider that.
Paul
wbmw, you have no reason to assume that it is a CPU problem. At 1.8GHz, the AGP would be at 75MHz, which causes many cards to have problems. Besides, we don't know anymore at this point about the rating of the RAM, other than being DDR333. We'll know that in a few days, though, I'm sure. Patience, grasshopper!
Paul
wbmw, nothing remarkable about it as far as I know, although I do have some questions about it, posed here:
http://www.investorshub.com/boards/read_msg.asp?message_id=1144008
Paul
EP & other process experts - What would be the economic advantages/disadvantages of continuing to run Tbred and Barton concurrently as opposed to doing Barton and Thorton which, from looking at the image, has to either have defective cache disabled or just intentionally disabled cache? Would the cost of the die size for Thorton be greater the cost savings of not having to run the separate die for Tbred? Just assume for the sake of argument that cache defects are not a substantial issue, and that the huge majority of the Thortons would have cache disabled for marketing/economic reasons rather than for defects? Would it make any sense?
Paul
Thorton sighting.
http://www.hexus.net/
wbmw & Tenchu, I think yb is looking at the wrong type of games. Q3 is known not to be very CPU dependent. Now some of the flight simulators and some games like Battlefield1942 put a good bit of stress on the CPU. While they run well now on the latest CPUs, if they make improvements in the AI of the games, that will add a huge hit on the CPU. If they plan that in such a way as to be able to take advantage of dual CPUs it can really smooth things out. As a matter of fact, one of the comments of many of the folks who have gone the dual route is not about raw speed increase, but in the lack of any hesitations and bogging down; "creamy smoothness" is one description I remember reading about.
Paul
Semi, maybe his lack of patience has to do with his attitude about the "lesser quality" people on this board. He frequently belittles this board over on SI.
Paul
"First Look" at nForce3 SK8N. Interesting to see that, in BIOS at least, the CPU temp was the same as the MB temp(33C, 91F). Bios supports easy overclocking, too, though I didn't notice if CPU voltage was adjustable.
http://www.amdzone.com/articleview.cfm?articleid=1298
EP, is Alzheimer's setting in? ;) You responded here http://www.investorshub.com/boards/read_msg.asp?message_id=1134227
to my post here http://www.investorshub.com/boards/read_msg.asp?message_id=1134164
although I noticed the ihub board clock seems to be all screwed up, because the post time is at least 8 hrs earlier than the actual time mine was posted. Or were you just referring to the past tense as opposed to the future tense?
Paul
jhalada, maybe his other nickname is "Slick Willy"?
EP, it's got to be either mood-altering drugs or sheer desperation. If things don't turn around almost immediately, I suspect there'll be another little niceguy sabbatical.
Paul
chipguy, nope, just saying you were making an apples and oranges comparison. ;)
chipguy, single processor in a 2P or 4P system would be in no way representative of the performance of a uniprocessor system.
Joe, great find! Having more than words to show from IBM is huge.
Paul
C'mon, EP, y'all have been harping on the lack of tier 1 OEM design wins, and now that one is going to be shown, it doesn't matter? A little consistency, please! Volume is not the only thing that is important, an IBM win is a large step toward being taken seriously and expanding into other tier 1s.
Paul
jhalada, why didn't you post this here, just on SI? Big news!
http://news.com.com/2100-1010_3-1020033.html?tag=fd_top
wbmw, yes, P4 has a legitimate win in SuperPi now, with a few caveats to at least consider for the real world. For one, the MHz gap required at the top end to beat XP is 1200MHz, which is more than the gap that exists now in available models. Second, look at the FSBs and tell me which is closer to a real world realistic FSB, the XP's(max of @20% above highest supported FSB speed) or the P4's (max of @ 60% above highest supported FSB speed). I will give Intel congrats on the job well done with the chipset, though. It would be interesting to compare with a similar FSB increase. But if wishes were horses, we'd all take a ride....
Paul
http://www16.big.or.jp/~bunnywk/cgi-bin/superpi.cgi
sgolds, yes, it's good to see the timeframes for beta/RC/retail reconfirmed.
Paul
wbmw, that's an invalid assumption, since A4 will be clocked higher than Opteron. How that all shakes out remains to be seen.
Paul
It was almost exclusively on SI(and RB?), but since I'm not a member there, I brought it up here for Joe to see, since I knew he was interested.
Paul
edgarcayce, at this point, I imagine that if you work with a vendor like newegg.com, which maintains fresh stock, you'd be hard-pressed to find anything but TBred Bs anymore, unless maybe it's a 2200+, since folks have avoided them for that reason. Besides, you may even get a later DLT3C version, like this guy did(link), who knows?
Paul
http://www.computing.net/cpus/wwwboard/forum/6020.html
yb, RTCW = Return to Castle Wolfenstein. eom
Looks like Jozef may be getting his wish.
Paul
http://pc.watch.impress.co.jp/docs/2003/0619/kaigai01l.gif
I wonder if this means that September-December will bring a faster A64 than 3400+, or just that they have slid the introduction forward?
Paul
http://www.amdzone.com/#5
wbmw, discussing QS is a dead horse that has been beaten into compost, that's why. There are many more relevant and interesting things to discuss.
Paul
wbmw, it is faster in some benchmarks, certainly not in most.
Paul
wbmw, a "smiley" might've helped. ;) I do have a sense of humor, I'm just sick of the wasted bandwidth on QS discussions.
Paul