Forums WoW Modding Support Archives WoWModding Support Archives [DiscordArchive] Eh, I'm pretty sure your machine is just overkill for the software. Have you tried the other MPQ too

[DiscordArchive] Eh, I'm pretty sure your machine is just overkill for the software. Have you tried the other MPQ too

[DiscordArchive] Eh, I'm pretty sure your machine is just overkill for the software. Have you tried the other MPQ too

rektbyfaith
Administrator
0
01-16-2025, 05:17 PM
#1
Archived author: Natrist • Posted: 2025-01-16T17:17:26.294000+00:00
Original source

Eh, I'm pretty sure your machine is just overkill for the software. Have you tried the other MPQ tools for comparison?
rektbyfaith
01-16-2025, 05:17 PM #1

Archived author: Natrist • Posted: 2025-01-16T17:17:26.294000+00:00
Original source

Eh, I'm pretty sure your machine is just overkill for the software. Have you tried the other MPQ tools for comparison?

rektbyfaith
Administrator
0
01-16-2025, 05:19 PM
#2
Archived author: Natrist • Posted: 2025-01-16T17:19:13.748000+00:00
Original source

It's amazing how far computer technology has come that you can load up a ~4GB archive in 12 seconds when it used to take at least a minute or more not even 10 years ago.
rektbyfaith
01-16-2025, 05:19 PM #2

Archived author: Natrist • Posted: 2025-01-16T17:19:13.748000+00:00
Original source

It's amazing how far computer technology has come that you can load up a ~4GB archive in 12 seconds when it used to take at least a minute or more not even 10 years ago.

rektbyfaith
Administrator
0
01-16-2025, 05:20 PM
#3
Archived author: Natrist • Posted: 2025-01-16T17:20:04.602000+00:00
Original source

I was looking at old WoW code the other day and they had so many optimizations and hacks around CD-ROM streaming and what not because of hardware limitations at the time.
rektbyfaith
01-16-2025, 05:20 PM #3

Archived author: Natrist • Posted: 2025-01-16T17:20:04.602000+00:00
Original source

I was looking at old WoW code the other day and they had so many optimizations and hacks around CD-ROM streaming and what not because of hardware limitations at the time.

rektbyfaith
Administrator
0
01-16-2025, 05:20 PM
#4
Archived author: Natrist • Posted: 2025-01-16T17:20:36.365000+00:00
Original source

Here you are flexing your 69 core computer lol.
rektbyfaith
01-16-2025, 05:20 PM #4

Archived author: Natrist • Posted: 2025-01-16T17:20:36.365000+00:00
Original source

Here you are flexing your 69 core computer lol.

rektbyfaith
Administrator
0
01-16-2025, 05:21 PM
#5
Archived author: Natrist • Posted: 2025-01-16T17:21:35.545000+00:00
Original source

I still think there's room for optimization in everything. It's shameful how some programmers give zero thought to it and instead rely on the hardware being overly powerful.
rektbyfaith
01-16-2025, 05:21 PM #5

Archived author: Natrist • Posted: 2025-01-16T17:21:35.545000+00:00
Original source

I still think there's room for optimization in everything. It's shameful how some programmers give zero thought to it and instead rely on the hardware being overly powerful.

rektbyfaith
Administrator
0
01-16-2025, 05:41 PM
#6
Archived author: 4bhorrent • Posted: 2025-01-16T17:41:39.753000+00:00
Original source

comparing to other tools is not as easy, bc i have to compare the very same operation.
mpq editor took quite some time to extract, but i am very sure that they do not multithread, or not to the same extent if it does.
also extracting to disk introduces the io bottleneck.
gonna do a extract to disk comparison one day, so it's easier.

right now my own lib bottlenecked by io, bc each file being loaded to ram is reading the compressed data for that file as a whole.
this can be optimized by loading everything at once and then just slicing into that array.
my machine has 64GiB, so no need to be too frugal lol
gonna implement a thin and obese mode, so one can choose how it performs.
rektbyfaith
01-16-2025, 05:41 PM #6

Archived author: 4bhorrent • Posted: 2025-01-16T17:41:39.753000+00:00
Original source

comparing to other tools is not as easy, bc i have to compare the very same operation.
mpq editor took quite some time to extract, but i am very sure that they do not multithread, or not to the same extent if it does.
also extracting to disk introduces the io bottleneck.
gonna do a extract to disk comparison one day, so it's easier.

right now my own lib bottlenecked by io, bc each file being loaded to ram is reading the compressed data for that file as a whole.
this can be optimized by loading everything at once and then just slicing into that array.
my machine has 64GiB, so no need to be too frugal lol
gonna implement a thin and obese mode, so one can choose how it performs.

rektbyfaith
Administrator
0
01-16-2025, 05:45 PM
#7
Archived author: 4bhorrent • Posted: 2025-01-16T17:45:32.018000+00:00
Original source

absolutely, and it's very nice that software slowly catches up to the state of hardware, in terms of multithreading and vector instructions.

i am always fascinated when i watch demos about how they solved programming tasks in old consoles or pc games, doing wild shenanigans which blow your mind, while still being inside the confines of the hardware.
rektbyfaith
01-16-2025, 05:45 PM #7

Archived author: 4bhorrent • Posted: 2025-01-16T17:45:32.018000+00:00
Original source

absolutely, and it's very nice that software slowly catches up to the state of hardware, in terms of multithreading and vector instructions.

i am always fascinated when i watch demos about how they solved programming tasks in old consoles or pc games, doing wild shenanigans which blow your mind, while still being inside the confines of the hardware.

rektbyfaith
Administrator
0
01-16-2025, 05:49 PM
#8
Archived author: 4bhorrent • Posted: 2025-01-16T17:49:54.671000+00:00
Original source

i find optimizing is the most fun port of programming tbh.
my lib takes advantage of the 64 bit register size while searching for a file, which cuts down on instructions while combing the hash table.
the optimized variant computes multiple hashes at once (you need 3 different hashes for searching a file) to cut down on overhead.
and this can even be further optimized by using SSE instructions to vectorize even beyond that
rektbyfaith
01-16-2025, 05:49 PM #8

Archived author: 4bhorrent • Posted: 2025-01-16T17:49:54.671000+00:00
Original source

i find optimizing is the most fun port of programming tbh.
my lib takes advantage of the 64 bit register size while searching for a file, which cuts down on instructions while combing the hash table.
the optimized variant computes multiple hashes at once (you need 3 different hashes for searching a file) to cut down on overhead.
and this can even be further optimized by using SSE instructions to vectorize even beyond that

rektbyfaith
Administrator
0
01-16-2025, 05:50 PM
#9
Archived author: 4bhorrent • Posted: 2025-01-16T17:50:58.918000+00:00
Original source

i am doing that btw bc i am still on that big client archive with 975 GiB of MPQ data, so it makes a difference for my use case lol
rektbyfaith
01-16-2025, 05:50 PM #9

Archived author: 4bhorrent • Posted: 2025-01-16T17:50:58.918000+00:00
Original source

i am doing that btw bc i am still on that big client archive with 975 GiB of MPQ data, so it makes a difference for my use case lol

Recently Browsing
 1 Guest(s)
Recently Browsing
 1 Guest(s)