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: kash johal who wrote (50770)	2/23/1999 7:53:00 PM
From: Tenchusatsu	Read Replies (1) \| Respond to of 1572161

Kash and all, Tom's latest CPU benchmarks in his article entitled "New CPUs from AMD and Intel": www5.tomshardware.com Some interesting highlights from the article: - The K6-3 450 MHz scores a little higher than a Pentium III 500 MHz on Business Winstone 99 under Windows 98. However, under Windows NT, the K6-3 runs slightly slower than a Pentium II 450 MHz. In Highend Winstone 99 under Windows NT (you can't run Highend under Windows 98), the K6-3 450 is slower than a Celeron 400. - Tom includes a graph showing the benefits of L3 cache in K6-3 systems. The scores are 100% for 0 L3, 103.6% for 512K L3, 105.6% for 1MB L3, and 108.2% for 2MB L3. (I believe all the other tests on the K6-3 were run with 1MB L3 cache.) - Tom also throws in a bunch of benchmarks for SSE-optimized software, but he does caution that these benchmark programs come from Intel. - Overall, Tom was pretty underwhelmed by the K6-3. He wasn't underwhelmed by the Pentium III, but then again, he wasn't expecting much in the first place. Tenchusatsu

To: kash johal who wrote (50770)	2/23/1999 11:48:00 PM
From: Petz	Read Replies (2) \| Respond to of 1572161

Winstone 99 IS the most important benchmark, even for "high end" users. For example, searching a 300 page HTML document, recalculating a 32K row by 80 column spreadsheet, searching a 100 Meg INBOX file are the kinds of things we actually have to wait for our computer to complete. All of these things are typically done using one of the three office suites used in the Winstone 99 benchmark. So to say that Winstone '99 is "one benchmark" is really inaccurate, its an amalgam of about 16 programs. Doing artistic work with Adobe Photoshop (included in "high end" benchmarks) would be fun, but not many people do this in an office setting. If the high-end benchmarks were broken down by application the results would be more useful. (In my case, I'd look at compile times and ignore the rest.) I also do a lot of simulations using my own code. 75% of these are integer and 25% floating point, so a word processor/spreadsheet based simulation is a good proxy for my 20K lines of C code. Petz