Recent Posts

Pages: [1] 2 3 ... 10
1
The Workshop / Re: Using Structures in Assembly
« Last post by riversr54 on Today at 08:41:38 AM »
I tried add a PTR in the CoorGrp initialization to specify exactly what the variable is but that didn't seem to make any difference. The second line below does in fact assembly but that is not what I really want to do.

Code: [Select]
Grp1 CoorGrp <COOR PTR testa, COOR PTR testb, COOR PTR testc>
Grp2 CoorGrp <<1,1>,<2,2>,<3,3>>

2
The Workshop / Using Structures in Assembly
« Last post by riversr54 on Today at 08:29:43 AM »
I'm trying to use structures within structures and am having some problems with initialization. As you can see below, I've created two structs. The first one (COOR) is a simple coordinate structure that contains two bytes, row and col. The second structure one (CoorGrp) is intended to contain three of the COOR structures. I am able to initialize COOR structures and use them successfully but I can't seem to find the right syntax to initialize one of the CoorGrp structures. I created three COOR variables, testa, testb and testb and then tried to add them to the CoorGrp with no luck. My ultimate goal is to create an array of the CoorGrp structures but that seems impossible given the problem that I am having with creating a single CoorGrp variable. The ERROR that if get with this code is: "Error A2181: initializer must be a string or single item" on both the lines where I try to initialize a CoorGrp variable.

I would really appreciate any help/suggestions.


Code: [Select]
;x/y coordinate structure
COOR STRUCT
row BYTE ?
col BYTE ?
COOR ENDS
CoorSize EQU 2

;group of three COORs structure
CoorGrp STRUCT
CoorA COOR <>
CoorB COOR <>
CoorC COOR <>
CoorGrp ENDS

;declare three COOR variables
testa COOR <65,65>
testb COOR <66,66>
testc COOR <67,67>

;declare a single CoorGrp variable
Grp1 CoorGrp <testa, testb, testc>

;declare an array of 2 CoorGrp elements
CoorGroups CoorGrp <testa, testb, testc>, <testa, testb, testc>

3
Romper Room / Re: Beta testers needed
« Last post by jj2007 on Today at 06:45:37 AM »
No box, no line... :rolleyes:

Code: [Select]
This is Windows version 10.0, build 18362
The long story: Windows version 10.0 1903, build 18362
running on Intel(R) Celeron(R) CPU  N2840  @ 2.16GHz   
4
The Campus / how do I start make very simple assembler?
« Last post by daydreamer on Today at 06:12:45 AM »
With parse strings and combine opcodes
Maybe simplest asm can output. Com files directly?
5
The Workshop / Re: N-Queen Problem
« Last post by LiaoMi on Today at 06:04:25 AM »
I had the opportunity to solve the 20-queens problem using a Tesla V100. The time is about 1/3 of what I got previously. But a Tesla costs a fortune.  :sad:





Yeah, GPU-Z is broken, does not report that CUDA is available.

Hi AW  :thumbsup:,

what an amazing power, I hope soon such video cards will be available to ordinary users.
6
Romper Room / Re: Beta testers needed
« Last post by LiaoMi on Today at 05:31:04 AM »
00406CAD - CreatePolygonRgn - GDI Leak - HRGN object Leak

Thanks a lot! So it's CreatePolyPolygonRgn, not CreatePolygonRgn. Yes, I do have an invoke CreatePolyPolygonRgn, but it never fails, and the handles (one per country) get deleted via ExitProcess. This is by design, otherwise neither tooltips nor "tell me which country I clicked on" would work. Is that why DeLeaker complains? Task Manager does not report handle leaks.

Quote
The line always appears, I tried 6 times, it appears if you expand the first graph completely, then reduce the inner window, and then move left, right, the second attempt already has a line  :angelic:

That one is a mystery, I can't see it. Which Windows version, and which screen resolution?

PolyPoly  :biggrin: Yes, the analyzer says there is a leak. .text:00406C99 call DeleteObject - called with zero, push dword ptr [esi] - there is always 0 in [esi] ?! It seemed to me that the last handler in ESI should be deleted, but if so, then there was a false alarm.

Even at the first start, the line is visible, when reducing the size it is visible ...
https://i.imgur.com/pVO4vgX.png

I have windows 10 RS6 1909 ...

The box in the upper left corner is fixed.

I still have it, only in a different form  :undecided: it is now always convex  :biggrin:
7
The Workshop / Re: Implementing QuickSort
« Last post by jj2007 on Today at 05:21:54 AM »
I had a quick look at qsort.asm - a detailed look is way beyond the time I can invest. What strikes my eye are numerous finit instructions. They are utterly slow. To give you an idea: One finit is over three times slower than measuring the length of a 100 byte string.

You should replace these finits with proper management of the fpu, i.e. ffree st(7), fstp st etc; where you consider finit essential, move it at least out of innermost loops.
8
Romper Room / Re: Beta testers needed
« Last post by jj2007 on Today at 05:13:23 AM »
00406CAD - CreatePolygonRgn - GDI Leak - HRGN object Leak

Thanks a lot! So it's CreatePolyPolygonRgn, not CreatePolygonRgn. Yes, I do have an invoke CreatePolyPolygonRgn, but it never fails, and the handles (one per country) get finally deleted via ExitProcess. This is by design, otherwise neither tooltips nor "tell me which country I clicked on" would work. Is that why DeLeaker complains? Task Manager does not report handle leaks, even when I frenetically resize the map, thus creating hundreds of region handles.

Quote
The line always appears, I tried 6 times, it appears if you expand the first graph completely, then reduce the inner window, and then move left, right, the second attempt already has a line  :angelic:

That one is a mystery, I can't see it. Which Windows version, and which screen resolution?
9
The Workshop / Re: N-Queen Problem
« Last post by AW on Today at 05:03:17 AM »
My earlier version of GPUZ (1.16.0) reports on my now aging NVIDIA GeForce GTX 960 that OpenCL, CUDA, Physx and DirectCompute 5.0 are present.

It is strange, no reason for GPUZ to be intimidated by a Tesla, a small program included in the CUDA SDK demos detects the Tesla the same way it detects others.   :rolleyes:
10
The Workshop / Implementing QuickSort
« Last post by UniverseIsASimulation on Today at 04:30:41 AM »
Hey, guys!
I am currently studying computer science and electrical engineering at the FERIT university in Osijek. So, for a seminar in algorithms and data structures, I've decided to attempt to implement QuickSort in my own simple low-level programming language, called AEC. You can see my work here:
https://github.com/FlatAssembler/ArithmeticExpressionCompiler/tree/master/QuickSort
So, the source code, written in AEC, is available here:
https://github.com/FlatAssembler/ArithmeticExpressionCompiler/blob/master/QuickSort/qsort.aec
The assembly my compiler (available for download here) produced is available here:
https://github.com/FlatAssembler/ArithmeticExpressionCompiler/blob/master/QuickSort/qsort.asm
The executable that FlatAssembler produced is available here:
https://github.com/FlatAssembler/ArithmeticExpressionCompiler/raw/master/QuickSort/qsort.exe
So, I've noticed that program is significantly slower than the programs produced by GCC.
Any idea how I can make it better?
Pages: [1] 2 3 ... 10