Re: Fastcode CharPosIEx B&V 1.0.4

Hi

Have a look at Northwood results

Version 1.0.3

CharPosIEx_JOH_IA32_3_d 4 501 663 1164
CharPosIEx_JOH_IA32_4_d 4 525 669 1194
CharPosIEx_DKC_IA32_26_a 4 550 697 1247

Version 1.0.4

CharPosIEx_DKC_IA32_26_a 4 549 711 1260
CharPosIEx_JOH_IA32_3_a C 512 758 1270
CharPosIEx_JOH_IA32_4_a C 528 757 1285

John's functions suddenly run slower. Can we find an explanation?

Best regards
Dennis Kjaer Christensen


.

Relevant Pages

  • Re: sparse / spalloc slower than pre-declaring a regular array
    ... some further explanation regarding why both of my methods are so much ... Not only does this appear to be slower than ... How would linear indexing into a matrix declared with spalloc do ...
    (comp.soft-sys.matlab)
  • Re: Fastcode AnsiStringReplace B&V 1.6.8
    ... Can you double check of the Northwood results. ... Best regards ... Jeg beskyttes af den gratis SPAMfighter til privatbrugere. ...
    (borland.public.delphi.language.basm)
  • Re: My own small benchmark
    ... Hans-Peter Diettrich schrieb: ... but the graphics are almost unreadable :-( ... with regards to my old K6-300. ... will run much slower on older machines. ...
    (comp.lang.pascal.delphi.misc)
  • Re: Fastcode Memory Manager Testing Status
    ... Dennis wrote: ... > I do not have the Sempron anymore and will have to complete WinME on ... Then you'll have to do the full suite of tests on the Northwood running ... Regards, ...
    (borland.public.delphi.language.basm)
  • Re: Looking for a list of processor instructions
    ... Are there any enhancements (except for clockrate) at all in it? ... Prescott tends to be slower than Northwood ...
    (comp.lang.asm.x86)