MASM32 Version 10 Posted.

[zemtex]

I'm not familiar with slow load times of notepad++. I use two western digital velociraptor disks in raid 0 perhaps thats why. It just pops up immediately when i click the icon.

[jj2007]

I'm not familiar with slow load times of notepad++. I use two western digital velociraptor disks in raid 0 perhaps thats why. It just pops up immediately when i click the icon.

For me it's two seconds on first load, one second on second etc load. Not a big deal.

[fearless]

Will we see an update of the MASM32 library in future that will incorporate some of the algo's tested over say the last 12 months? I'll admit that at times the testing of algos, while impressive, leaves me confused as to which ones to implement. There are a lot of string related algos: comparing, searching, length, zeroing etc, that all have merits. Tbh i like the idea of a library that is updated and incorporates the best (best is hard to qualify of course, best in terms of speed, size or something in between?) of them.

[jj2007]

speed, size or something in between?

Lamps, of course ;-)

Hutch seems a bit conservative on this topic, and rightly so: Many algos have not been extensively tested, and may present problems under not so obvious conditions.

Another question to discuss is the introduction of SSE2 algos. They are a lot faster, but some old hardware would drop out...

[dedndave]

i am sure he's working on it   :bg

updating this thing is a bit of a monster, with the advent of win 7
and - i guess he has something going on for 64-bit, too - i can imagine what a mess that is

unfortunately, the masm32 package isn't his only thing in life - lol
last i knew, he was doing some remodeling
he has hardware toys to play with
and, the crocodiles in the outback demand his time   :bdg

we can either have frequently updated libraries, or an administrator that has a social life
i guess the latter is best

you can, however, put some of the new algos into your own library for use
you may not always want to alter the masm32 libraries, as it may make you incompatible with other members

[hutch--]

it does tend to be a workload issue, at times you can get something useful from the lab algo testing but often you end up with a pile of incompatible inconclusive results. Anything of late that gets added to the library must pass exhaustive testing as I have been caught in the past with contributed algos that broke in some contexts.

RE: The use of SSE, its a problem of the higher you go, the less people can use it. I would prefer to see an SSE library where the end user knows what is required to run it but the masm32 library is intentionally backwards compatible with integer and FP instructions only.

[dedndave]

well - i think if the programmer wants SSE - they can implement it on a per-install basis or something, replacing the existing algos as required
compatibility is nice to have and a program only requires selected algos to be super-fast

[Rockphorr]

RE: The use of SSE, its a problem of the higher you go, the less people can use it. I would prefer to see an SSE library where the end user knows what is required to run it but the masm32 library is intentionally backwards compatible with integer and FP instructions only.

When will we can see  SSE library in package ??

[hutch--]

When you feel like writing one and documenting it for each processor.

[xandaz]

   Does anyone know if hutch posts updated versions of the masm package or if he leaves it untouched until a new version package is released? I have my masm10 for a while and was wondering if it would be about time to download it again.
   Thanks and bests
   X

[Rockphorr]

When you feel like writing one and documenting it for each processor.

I think that programmers team make masm32. You are right - work alone is hard.