Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - Dariusz Piatkowski

Pages: 1 2 [3] 4 5 ... 76
31
Programming / Re: i686 vs Pentium 4
« on: March 08, 2022, 02:04:27 pm »
Dave, everyone...

So anyways, my next idea is to benchmark the differences in optimizing for different CPU architectures. I found this, https://www.eembc.org/coremark/index.php a benchmark designed to test integer and pipeline speed on various CPU's see the Details part of the above page.
So I downloaded the source and built it and now share it, it is attached and includes the logs I generated...

Nice!!! Thank you...

Here are the results for my Phenom II X6 running at 3.8GHz:

Code: [Select]
2K performance run parameters for coremark.
CoreMark Size    : 666
Total ticks      : 17587
Total time (secs): 17.587000
Iterations/Sec   : 17058.054245
Iterations       : 300000
Compiler version : GCC9.2.0 20190812 (OS/2 RPM build 9.2.0-5.oc00)
Compiler flags   : -O2 -DPERFORMANCE_RUN=1
Memory location  : Please put data memory location here
                        (e.g. code in flash, data on heap etc)
seedcrc          : 0xe9f5
[0]crclist       : 0xe714
[0]crcmatrix     : 0x1fd7
[0]crcstate      : 0x8e3a
[0]crcfinal      : 0xcc42
Correct operation validated. See README.md for run and reporting rules.
CoreMark 1.0 : 17058.054245 / GCC9.2.0 20190812 (OS/2 RPM build 9.2.0-5.oc00) -O
2 -DPERFORMANCE_RUN=1   / Heap

I'll play around with the different optimization flags Dave which you discussed in your post next. Also, the above is a result when executed on my system in as-is state, so various CPU spikes, etc...not quieted by any means! lol

32
Programming / Re: i686 vs Pentium 4
« on: March 08, 2022, 03:22:50 am »
As I stated at the top, the conversation should be about the architecture, not the benefits/drawbacks of RPM, that ship has sailed. Our volunteers are comfortable with RPM and we get RPM and we do need a package manager, which Warpin is not...

Thanks Dave fore re-focusing this discussion on the main topic.

...The question is how to easily fix the performance hit from using Netburst on most anything that isn't on Netburst...

So what do we actually know, as opposed to believe, this performance penalty to be?

As you know I'm happily chugging away using my AMD Phenom II CPU, old tech by today's standards, but it does support some of the instructions that have the potential to improve system performance, that being: SSE, SSE2, SSE4a. So from that perspective I do not see anything wrong with the move towards pentium4 releases.

If anything, I wish we actually had a more real choice, that being CPU specific package builds, basically along the lines of what you did for me building my Phenom specific FF release. In that case, I did genuinely see an improved performance and could benchmark enough to actually put a real metric on the results.

33
Hmm...interesting because I have been using Thunderbird here with my Gmail account and had the 2-way authentication turned on quite some time ago.

Google does have a process in place (application password) that will allow you to define app-specific password that the application can authenticate with. Works quite fine with TB, I only wish the Google Contacts add-on would handle that...

Can't speak to PMMail, but I suppose that should apply as well given that you need to provide ID/PASS credentials there as well.

34
Hekki,

I can use my internet bank with Simple Browser. It is slow and crashes often, but it doesn't leak shared memory like FreeRDP.

Have you ever tried troublshooting that FreeRDP memory leak?

I have and I suspect due to lack of know-how wasn't able to produce real results. But yes, I do see this here and it's the only -ve to that approach.

35
Dave,

...Hi Dariusz, we don't all have a spare Windows or Linux machine to use...

You know, I had a crazy idea at lunch today on this very topic: I should find the smallest factor Winx hardware box, so that it basically can work as a headless 'app server' for the sole purpose of providing the kind of app functionality we have been missing on our platform for quite some time, and most likely will never see again.

Browser is a perfect example of this, and while I personally prefer FF, heck, others may want to use Chrome.

...As for funding, I don't think there are any big fish who care about having a browser left besides Arca Noae and I had the impression they and Bitwise don't get along as well as they used to, not that I really know...

Well, this "big fish" reference, while I think today that may in fact exclusively apply to AN, was meant to apply to any business that still has the need for updated OS/2 browser functionality.

I make no assumptions about what our OS/2 platform looks like today...nor am I particularly optimistic that out there is some organization that has deep pockets and wishes to continue doing what all of us have been doing. But does anyone today actually know the true picture of our ENTIRE OS/2 installed base?

Dave, I do agree with you re: AN's direction, or perhaps lack thereof. Having said that, they have done some incredible things with AOS that it's hard to 'blame' them for lack of browser strategy...although I think everyone will agree that it is a very questionable position to be in.

36
ivan, everyone...

...If it won't do those things I will continue using Dave's Firefox ESR 38.8.0 on OS/2 or the Linux Firefox ESR via my KVM switch.

You bring up an excellent point, and even though they say "you don't look a gift horse in the mouth", the reality is that by now, after some years of use, most of us have established a browser platform for interacting with the WWW.

So while my FF 45.9 may not be capable of successfully rendering ALL the sites out there, the reality of my life is that I have far too much of a 'FF 45.9 configured' setup to simply move on over to a browser that only does the basic rendering for ALL of the pages out there.

In fact, as an alternative it is FAR cheaper and FAR more robust to utilize RDP access to a Win/Linux based browser install, or the KVM switch that you ivan are suggesting.

Here is what's been eating away at me: the funding drive, while generally a great and just effort, seems to be single handedly relying on us, the OS/2 end-users to cover the costs. So where are the big-fish of our platform in all of this? Does anyone know? Have those kind of conversations taken place and have any results been arrived at?

If you want to talk full-disclosure, well, let's "talk the talk, and walk the walk"...much easier to raise additional funds if we all understand the full playing field!

37
Programming / Re: Ghostscript - 9.18 build, any advice?
« on: February 09, 2022, 03:48:18 am »
Dave,

Maybe download the src rpm, unrpm it and take a look at how it was built, libs, flags etc.

I grabbed the source ZIP file that BWW has on their GitHub project page...what is the "src rpm", do you mean to say actual RPM package that has the full source?

I see they have a Git https URL, and while I do have Git installed here, I haven't done any real project work with it, so I'd rather not mix up two learning curves at the same time! LOL

38
Programming / Ghostscript - 9.18 build, any advice?
« on: February 09, 2022, 02:47:24 am »
 ;)

Alright, so my help ticket hasn't resulted in an answer from the official BWW channel: it sounds like Silvan is alone/himself handling these, and since it's a volunteer effort I best make plans for an alternative approach to debugging my "Failed to print with GS 9.18". Just our reality I think, no fault of his.

Soo....I downloaded BWW's GS codebase, deployed locally, I have studied the source for a couple of days now. Did my share of the official documentation review...I have GCC as well as IBMCPP available here and it looks like the OS2.MAK file can handle either one.

Therefore, before I take the plunge and run with it: any last-minute advice as to what to watch out for???

It's a big app, numerous pieces (display and print devices, etc, etc), but the makefile itself doesn't actually look all that daunting.

Thanks!

39
Setup & Installation / Re: OS/2 as it was
« on: February 09, 2022, 02:14:22 am »
Hi ivan,

...My main work version of OS/2 is based on ArcaOS 5.6 with most of the linux junk stripped out.  It is connected to a hanns.g 28 inch monitor via an hdmi kvm switch, my linux box is connected to that monitor as well and is mainly used for firefox to allow me to get to some sites that require a newer version to work...

Consider deploying FreeRDP client on OS/2 and get the server side up on the Linux box and use the seamless window RDP connection to put a Linux FF right on your OS/2 Desktop.

I have been taking that approach with my Win 7/10 based apps for a little over a year now...and actually use a fulls-creen RDP connection in about 90% of the cases as I usually need to have numerous other Win apps besides FF up and running.

40
Setup & Installation / Re: SSD partition miss aligned.
« on: February 05, 2022, 04:54:29 pm »
Remy, everyone...

It's the usual problem, need to do all partitioning with OS/2 aware tools, and do some math to have the 4k sector partitioning, basically JFS 4k sectors need to align with the SSD's 4k sectors...
...
Before, I had H255 and S63  (of course, sector size always left at default 512...)
But this resulted to unaligned partition while Dfsee found it aligned (ok for HDD) but it isn't for SSD.
...

Dave brought up an important point, at least from a performance perspective: the alignment of hardware and software implementation.

So 4k sector size for SSD, and unless you MUST pick something different from JFS IFS perspective, you would want the same matching 4k JFS blocksize (see format's /BS parameter).

The HELP page for our format specifically states:

Quote
...The block size specified must not be smaller than the sector size of the device....

...however what I do not know (having never tried it myself) is what would happen if you do try a JFS blocksize <4K?

Anyways, re: the above sync, my expectation would be that this would assure a minimal overhead for all processing.

41
Utilities / Re: MeShell commandline frontend timetable
« on: February 03, 2022, 03:07:23 pm »
Hi Martin,

The behaviour is how I expected. If you type "mode 60,25", then the VIO window gets bigger and you will get the behaviour you want...

Hmm, I am sorry but I have to disagree with you on that.

My expectation is that the commands I issue within MeShell apply to the user (front-facing) interface (MeShell). As such, given that this interface is the MeShell window, I am expecting that MeShell is creating a comparable VIO background processing window.

In other words: if my MeShell window is defined as 80x60, and I issue a "dir /p" command, your code should make sure that the background VIO window executes with the matching criteria. Therefore, perhaps an enhancement to create a VIO window that physically matches the definition of MeShell would suffice?

What I am literally suggesting is:

1) let the user define MeShell window
2) MeShell itself should understand that definition (sizing is really what we are talking about here)
3) MeShell should then create the background VIO window with the matching sizing (yup, could be as simple as issuing that MODE command)
- after all, the benefit of using MeShell should be that I am not required to perform additional adjustments to the background VIO windows, otherwise if I do, then why not use a regular VIO window instead?

Is it worth to compile the conflicts between MeShell and Xit?

What do you mean by that?

Are you interested in troubleshooting the conflict? I do not think there is much value in doing that...I was able to put MeShell on Xit's Exception List, and that fixed it.

42
Setup & Installation / Re: SSD partition miss aligned.
« on: February 03, 2022, 04:12:11 am »
Andreas,

I use DFsee (or dfsanos2 via minilvm) to create partitions. All my SSDs (including Samsung 860 EVO, 500 GB) work.

OK, but how do you know that your partition boundaries are actually aligned?

Remy's point is that unless you have that 4K alignment, it is known that the performance of your SSD is degraded given multiple memory page reads that occur otherwise (I believe, but easy to find detailed on-line explanation out there).

43
Programming / Re: GSView - make complains of "No such file or directory"
« on: February 03, 2022, 04:09:22 am »
Hi Paul,

From the readme of my last gs build:
"4) GSView requires manual changes to the Ghostscript Include statement (Options->Advanced Configuration) - I used 'c:\gs\gs9.15\resource\init;C:\GS\gs9.15\lib;C:\GS\gs9.15\Resource\font;C:\GS\fonts;c:\psfonts;' to make it work
5) the pswrite device that GSView uses to print has been removed from Ghostscript 9.x. To print using ghostview, modify x:\os2\gvpm.ini and replace all cases of pswrite with ps2write"...

Oh, believe me, I went down that road long time ago. In fact you may not remember, but you had previously tried to helpe me with this (getting your port of GS9.15 printing) and we could not get it to work.

I have a SINGLE occurence of gvpm.ini file on my partition, any changes I make are always reflected in it. Yet, GSView still tosses an error.

That is what caused me to use GS9.04, until Lucide 1.5.0GA (which is about late 2019 I think) came out, so I switched to that nearly fulltime, but like I said earlier, sometimes I still have printing problems there.

Anyways, appreciate the response nonetheless!

44
Setup & Installation / Re: SSD partition miss aligned.
« on: February 03, 2022, 03:30:00 am »
It's the usual problem, need to do all partitioning with OS/2 aware tools, and do some math to have the 4k sector partitioning, basically JFS 4k sectors need to align with the SSD's 4k sectors...

But Dave, for those of us 'mere mortals' (when it comes to doing that mapping RIGHT and knowing how to instruct the tools to implement) this is a nearly hopeless situation.

Remy,
I struggled a bit through this a couple of years ago when I decided to move my OS/2 stuff to a SSD. I noticed that DFSEE did actually provide a way to build an OS/2 partition where the goal was to implement this very alignment. I tried to do that, but the resulting disk GEO was looking really strange and the remaining OS/2 disk tools complained.

Eventually, having spent a pile of time on this and NOT knowing whether all I've done did actually get me a right or wrong setup, I finally just partitioned with LVM like I've done before and off I went having spent far too much time just trying to get OS/2 to benefit from the latest tech.

Total frustration! So I feel your pain.

Unfortunately, I do not have an answer. For what it's worth I'm about to re-do all of that given that I wanted to move my stuff to a larger SSD, that being a Samsung 860Evo (from my current 850Evo drive). I will experiment again, but truth be told, as the tech has moved forward it's been harder and harder to keep all my trusty OS/2 stuff afloat, let alone actually being able to benefit from the better hardware.

I hate to mentally "toss in the towel", but I feel like that "end of the OS/2 road" may be nearer for me than I had previously wanted to believe, or perhaps even realized. Grrrhhh....I hate that feeling!

45
Applications / GSView and Ghostscript - printing, how?
« on: February 02, 2022, 03:24:14 pm »
We've got a separate discussion (Programming forum => https://www.os2world.com/forum/index.php/topic,2985.msg33707/boardseen.html) on the challenge I've encountered attempting to use GS 9.18 to faciliate printing with GSView.

As I researched this further, and discovered that perhaps some of the printing functionality is a little off in this release (as compared to earlier ones) that made me wonder about how the rest of you are printing documents that may utilize GS for rendering.

Specifically, and most often I suspect, that would apply to PDFs.

In my case, I can use Lucide to print PDFs. Occasionally I run into an issue though and the print job produces an error on the printer itself (the root cause is unknown to me, but it's some kind of a PS stack overflow and my prior investigation suggested this may be a printer driver issue, which as we all know simply isn't going to get fixed, so case closed).

In such cases, I would most often rely on GSView to handle the printing task, and that has been working quite fine with GS9.04, but we are at 9.18 now and GSView with GS9.04 started to show it's age (various rendering errors would show up here and there), and much slower than GS9.18.

Therefore, I am curious how you print from GSView with whatever version of GS?

With GS9.04 I use the 'pswrite' Device and have the 'PostScript Printer' check box marked given that my Brother HL-5470DW is a PS printer.

With GS9.18 I tried using 'ps2write' Device since 'pswrite' has been obsoleted, but the '\\spool\printer_name\' printer identifier no longer appears to print in GS9.18.

Thanks!

Pages: 1 2 [3] 4 5 ... 76