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.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Tenchusatsu who wrote (100479)	3/28/2000 3:49:00 PM
From: Ali Chen	Respond to of 1576349

<I don't consider those benchmarks to be valid, because an overclocked 440BX is running the SDRAM so far out of spec that it cannot be shipped in that manner..> What a pile of nonsense, my dear. The benchmarks are certainly valid as a proof of concept. And the concept was very simple: the PC-133 SDRAM with 133FSB system solution is SUPERIOR in performance relative to twice-in-bandwidth and 10x more expensive RAMBUS. This is an undisputable fact. And no one implied that the 440 chip is shippable as is, without little product re-engineering. Then, your "running SDRAM so far out of spec" is a total bull. First, the "spec" usually is written based on subsystem functionality with certain guardbands. Second, there is no real spec on PC-100 besides ultimate compatibility with the 440BX chipset. Third, there are numerous reports that the _current_ 440BX with _current_ crop of PC-133 is functional and well up to 170MHz, (see recent results on dot.com.jp sites). This gives plenty of margins for you to settle around 133 MHz. <They tried the same thing that Tom did with less success (in stability) than Tom claimed to achieve.> Too bad. Your co-workers prove themselves as ignorants in this case. I can tell you what to do to achieve enough stability: you need to cool down the 440 chip when running at elevated frequencies. Too bad your co-workers have little ideas about thermal management, as it was also proven in certain other system areas :) What you better to admit is that some short-sighted Intel manager has decided few years ago that there will be never a need to combine AGP-133 with FSB-133, and dropped the requirement of 1:1 FSB/AGP ratio in the 440 chip. By some luck, the chip has the PCI running out of independent clock, and no overclocking happens if the system uses a PCI video card. - Ali, the "screwer-driver"

To: Tenchusatsu who wrote (100479)	3/29/2000 2:33:00 AM
From: Joe NYC	Read Replies (2) \| Respond to of 1576349

[Overclocked 440BX vs. 820] I don't consider those benchmarks to be valid, because an overclocked 440BX is running the SDRAM so far out of spec that it cannot be shipped in that manner The 440BX was overclocked, but it is only because Intel chose not to release a version of BX that supports 133 MHz FSB / memory. Are you suggesting that is incapable of releasing 133 MHz BX equivalent? Tiny Via did it. As far as SDRAM running out of spec, you are right, running with CAS-2 is out of spec of most PC-133 RAM out there, but there was no demand until very recently for PC-133 of any variety, since there was no motherboard support for PC-133. Since this has changed, and there is a real market out there for PC-133 memory, it will not take long for CAS-2 to became widely available. If you recall, when PC-100 came out, CAS-2 variety was hard to get as well. I checked this with a few coworkers, who confirmed this for me. They tried the same thing that Tom did with less success (in stability) than Tom claimed to achieve. Are you saying that Intel employees are not capable of overclocking a system? The key is good memory, if you want CAS-2. BX chip can get hot, but you can replace heatsink with a heatsink with fan attached to bring the temperature down to normal. Joe