Reply 4560 of 4893, by H3nrik V!
- Rank
- Oldbie
Nexxen wrote on 2024-01-31, 00:00:The Ultimate 686 Benchmark Comparison […]
appiah4 wrote on 2024-01-30, 21:24:Umm, no? […]
H3nrik V! wrote on 2024-01-30, 17:00:Again, ALU wise, FPU is hopeless ...
Umm, no?
https://thandor.net/benchmark/33
PR133 is about Pentium 90 levels of performance in Quake (which is more than twice as fast as an Intel 486DX4-100), quite possibly the most FPU intensive application you will ever run on it.. So no, it is not hopeless and CERTAINLY NOT a 486 CPU.
I have a K5 PR133 in my 586 build, and I can tell you first hand that its FPU is decent, if you would believe me. There is this weird belief in the retro community that anything non-Intel on the Socket 7 has a terrible FPU. Probably has to do with Cyrix having a particularly weak FPU. Regardless, the K5 has a pretty good FPU and the K6 FPU is nearly on par with Intel..
The Ultimate 686 Benchmark Comparison
Always worth a look.
Here:
download/file.php?id=11657K5 133 is like 10-15% behind a P133 (ok, my math sucks but you get it).
Just reporting, no opinion given.
No, a 133 MHz K5 is 10-15% behind a P133; a K5-PR133 is more like 33% behind a P133 in FPU benchmart (40.1 vs. 61.4) making the P133 FPU in the ballpark of 50% faster. ALU vise, the difference is close to neglible (60.4 vs. 61.5); which is probably also why AMD used the PR-marking ... The 133 MHz K5 was the PR200 😀
If it's dual it's kind of cool ... 😎
--- GA586DX --- P2B-DS --- BP6 ---
Please use the "quote" option if asking questions to what I write - it will really up the chances of me noticing 😀