Tom, We agree .. I just can't write a book via email describing all of the nuances.
Even on a single CPU core the performance should be better, but I suspect what we have is a "bogomips" like benchmark which is good at measuring the effective clock rate, not the actual processing power of the thing being tested. Which was the point I was trying to make. Just as the chip has many CPU cores, each CPU core has much more parallelism built into it. So even individual cores are going to be capable of much more work. Which can't be exposed in a toy benchmark. Mike On Thu, Mar 23, 2023 at 1:37 PM tom ehlert <t...@drivesnapshot.de> wrote: > Dear Mike, > > I almost always agree with; not so in this case. > > it's not only the (performance per core) x (number of cores). > > the i5 is also far more capable then a K6-2 in single threaded/core mode - > as it would > be in any xxDOS setting. > > Tom > > > > > > am Donnerstag, 23. März 2023 um 20:42 schrieben Sie: > > > Your measurement methodology needs help? > > > > The i5 is capable of far more work than the K6-2-350. What you are > > doing is comparing a bus to a passenger car and stating that they > > both are effective at carrying one person, and therefore the bus has no > additional value. > > > > When running a simple single-threaded workload sure, the i5 might > > not look like it's worth the extra silicon. Now go run 8 or 16 or > > 32 copies of the same workload in parallel, and tell me how that > > goes for the K6-2. Modern chips are designed for very different > > things than chips from 20+ years ago, so the comparison is kind of > meaningless. > > > > You're effectively measuring the clock speed, not what the chip is > > capable of. And this isn't a surprise, but a 10x faster clock on a > > single threaded benchmark results in 10x performance. While > > completely ignoring the rest of the chip. We don't benchmark buses > against passenger cars. > > > > > > On Thu, Mar 23, 2023 at 12:33 PM Bret Johnson <bretj...@juno.com> wrote: > > >>> Even though a 350 MHz K6-2 is WAY faster than my 3.3 GHz i5? And > >>> again, I know it doesn't make any sense, but it's still true. > >>> > >>> As far as I can tell, the Emperor has no Clothes. > >> > >> I'm confused by this. You are claiming that a 350 Mhz K6-2 is "WAY" > faster than a 3.3 GHz i5? In what respect? > > > My K6-2-350 runs about 15 times faster than a 33 MHz 386, while my > > 3.3 GHz i5 runs only about 10 times faster, in both cases with the > > caches enabled. What do you not understand? > > > You can try to claim that the SLOWDOWN test I performed is somehow > > "tainted" or "unrealistic" or "unfair" or "sub-optimal" or something > > like that, but the results are very real and are simply what they > > are. Also, as already stated, the purpose of SLOWDOWN is not to be > > a benchmark but you can indirectly use it as a benchmark. > > > > _______________________________________________ > > Freedos-devel mailing list > > Freedos-devel@lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/freedos-devel > > > > > > > Mit freundlichen Grüßen / with kind regards > Tom Ehlert > +49-15151898538 > > > > _______________________________________________ > Freedos-devel mailing list > Freedos-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/freedos-devel >
_______________________________________________ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel
