ROS testing on real hardware

Here you can discuss ReactOS related topics.

Moderator: Moderator Team

Webunny
Posts: 1201
Joined: Sat Apr 28, 2012 1:30 pm

Re: ROS testing on real hardware

Post by Webunny »

oldman wrote:
Post by Webunny » 05 Dec 2013 20:26
let me know if your success-rate is higher than 25%; it may give an indication in ROS is slowly getting more compatible with real hardware.
If you are referring to Ros installing and running, then I have this year been getting something around 95% success rate and possibly 100%! But I only test on one computer.
The issues that I have, are more to do with software that does not install or does not run correctly. The only OS related issue that I can think of, is, that there is a problem with the log in, but that was still being coded.
No, I meant of different rigs. I tried out 4 different computers back when testing 3.14. Only this one (an intel) worked (the three others just didn't work). I think the others were all AMD's. But that can be coincidence, of course. Point is, the more different PC's are tested, the better. Even if they bork completely, it would still be great to see if the compatibility with (different kinds of) PC's are getting better (or not).

oldman
Posts: 1128
Joined: Sun Dec 20, 2009 1:23 pm

Re: ROS testing on real hardware

Post by oldman »

Webunny wrote:
No, I meant of different rigs. I tried out 4 different computers back when testing 3.14. Only this one (an intel) worked (the three others just didn't work). I think the others were all AMD's. But that can be coincidence, of course. Point is, the more different PC's are tested, the better. Even if they bork completely, it would still be great to see if the compatibility with (different kinds of) PC's are getting better (or not).
My test computer has an AMD Athlon 64 Processor 3200+ with nVidia nForce3 chipset.
It is second-hand (or even third-hand) and it's not badged, so must be from a small shop or home made.

Is there a hardware database for computers that work with Ros?
Please keep the Windows classic (9x/2000) look and feel.
The layman's guides to - debugging - bug reporting - compiling - ISO remaster.
They may help you with a problem, so do have a look at them.

Webunny
Posts: 1201
Joined: Sat Apr 28, 2012 1:30 pm

Re: ROS testing on real hardware

Post by Webunny »

oldman wrote:
Webunny wrote:
No, I meant of different rigs. I tried out 4 different computers back when testing 3.14. Only this one (an intel) worked (the three others just didn't work). I think the others were all AMD's. But that can be coincidence, of course. Point is, the more different PC's are tested, the better. Even if they bork completely, it would still be great to see if the compatibility with (different kinds of) PC's are getting better (or not).
My test computer has an AMD Athlon 64 Processor 3200+ with nVidia nForce3 chipset.
It is second-hand (or even third-hand) and it's not badged, so must be from a small shop or home made.

Is there a hardware database for computers that work with Ros?
I think there used to be one on the old site, but I'm not sure it's ported (yet). Or even if it will be. In theory, the more binary compatible Reactos gets, the more it should work with virtual all desktop PC's.

What I've noted with the latest versions, when I take the card out and get it too work that way, is that it now auto-detects my ethernetcard automatically. This is a huge improvement, as far as I'm concerned, and will certainly boost the user-friendliness in regard to the enduser-experience. Congratz and well done to whichever dev made/fixed that. Auto-search actually amounts to *something* for the first time. And I'm sure it will become better fast now, since all one has to do is add extra much-used drivers.

oldman
Posts: 1128
Joined: Sun Dec 20, 2009 1:23 pm

Re: ROS testing on real hardware

Post by oldman »

I have now installed revision 61208 (my latest down-load) on a old Fujitsu ErgoPro with a Pentium 2 processor, 96Mb ram, 3.2Gb hard drive (partitioned C:, D:), a PCI Ethernet card with a Realtek 8139c chip and a S3 Virge DX PCI video card. I have used this computer for testing in the past.
There was no hardware wizard when booting into the desktop, because Ros found the Ethernet card and used it's own drivers(I believe that Ros ships with a 8139 driver), also the vesa drivers are always good enough.
There, you have another computer to add to your list of working with ReactOS!
Please keep the Windows classic (9x/2000) look and feel.
The layman's guides to - debugging - bug reporting - compiling - ISO remaster.
They may help you with a problem, so do have a look at them.

Webunny
Posts: 1201
Joined: Sat Apr 28, 2012 1:30 pm

Re: ROS testing on real hardware

Post by Webunny »

oldman wrote:I have now installed revision 61208 (my latest down-load) on a old Fujitsu ErgoPro with a Pentium 2 processor, 96Mb ram, 3.2Gb hard drive (partitioned C:, D:), a PCI Ethernet card with a Realtek 8139c chip and a S3 Virge DX PCI video card. I have used this computer for testing in the past.
There was no hardware wizard when booting into the desktop, because Ros found the Ethernet card and used it's own drivers(I believe that Ros ships with a 8139 driver), also the vesa drivers are always good enough.
There, you have another computer to add to your list of working with ReactOS!
Now that you said it, I've been thinking about it, and maybe it wouldn't be bad to make a page specifically about hardware-that-works-with-ROS. A wikipage, for instance. There are other threads here of users/testers that tried it on real hardware, but they are always dispersed and not centrally managed (so one could find all confirmed working hardware at one glance in one central place). Alternatively, one could go for one central thread on this forum, but it seems it would advisable to use a sticky, in that case.

The wiki is a collaborative effort, so I think we can go ahead with it as long as we take it onto ourselves to maintain it (obviously, the devs have better things to do). Maybe as a (sub)link of 'supported hardware'', where we type the rig we have such as kind of PC, how much RAM, motherboard, etc.)? Would you (or others) be interested in such a thing?

mametoc
Posts: 173
Joined: Tue Nov 27, 2012 11:31 pm

Re: ROS testing on real hardware

Post by mametoc »

What criteria?: Official releases?, trunk builds?, both?, will be created a new one PC config entry becouse only differs from other config whit twice of RAM and a diferent videocard?. I dunno at what point this might be useful from an user and/or developer technical view.

Webunny
Posts: 1201
Joined: Sat Apr 28, 2012 1:30 pm

Re: ROS testing on real hardware

Post by Webunny »

mametoc wrote:What criteria?: Official releases?, trunk builds?, both?, will be created a new one PC config entry becouse only differs from other config whit twice of RAM and a diferent videocard?. I dunno at what point this might be useful from an user and/or developer technical view.
I don't see any insurmountable difficulties. One can just indicate the build one is using, for instance. The difference in RAM is hardly to be detrimental if it's going to be more, and yes, the videocard can play a role. All in all, seeing the current hardware that it's really tested on, it differs greatly, so I don't think one has to worry about it becoming meaningless differences (like RAM) any time soon. As far as usefulness: if people want to try out real hardware testing, at least they can have an idea what will work and what not.

In the end, usefulness is determined by people being interested in it. If enough people are interested to maintain such a page, why not do it?

mametoc
Posts: 173
Joined: Tue Nov 27, 2012 11:31 pm

Re: ROS testing on real hardware

Post by mametoc »

Well you know about "by people being interested in it" is not a suffice reason (like the mascot thing) but maybe someone can explain "ok, let's do it+quite reasons" or "directly not+few reasons".

fred02
Posts: 551
Joined: Thu Nov 22, 2007 5:54 pm

Re: ROS testing on real hardware

Post by fred02 »

Webunny wrote:
mametoc wrote:What criteria?: Official releases?, trunk builds?, both?, will be created a new one PC config entry becouse only differs from other config whit twice of RAM and a diferent videocard?. I dunno at what point this might be useful from an user and/or developer technical view.
I don't see any insurmountable difficulties. One can just indicate the build one is using, for instance. The difference in RAM is hardly to be detrimental if it's going to be more, and yes, the videocard can play a role. All in all, seeing the current hardware that it's really tested on, it differs greatly, so I don't think one has to worry about it becoming meaningless differences (like RAM) any time soon. As far as usefulness: if people want to try out real hardware testing, at least they can have an idea what will work and what not.

In the end, usefulness is determined by people being interested in it. If enough people are interested to maintain such a page, why not do it?
mametoc wrote:Well you know about "by people being interested in it" is not a suffice reason (like the mascot thing) but maybe someone can explain "ok, let's do it+quite reasons" or "directly not+few reasons".
There is already this page: Supported Hardware and I have memories of some people posting their (un)successful results on the forum, but can't find it right now.
As for maintenance: it is a Wiki, just edit it. ;)

mametoc
Posts: 173
Joined: Tue Nov 27, 2012 11:31 pm

Re: ROS testing on real hardware

Post by mametoc »

Webunny wrote:Now that you said it, I've been thinking about it, and maybe it wouldn't be bad to make a page specifically about hardware-that-works-with-ROS. A wikipage, for instance. There are other threads here of users/testers that tried it on real hardware, but they are always dispersed and not centrally managed (so one could find all confirmed working hardware at one glance in one central place). Alternatively, one could go for one central thread on this forum, but it seems it would advisable to use a sticky, in that case.

The wiki is a collaborative effort, so I think we can go ahead with it as long as we take it onto ourselves to maintain it (obviously, the devs have better things to do). Maybe as a (sub)link of 'supported hardware'', where we type the rig we have such as kind of PC, how much RAM, motherboard, etc.)? Would you (or others) be interested in such a thing?
He already known that but wants like an add list based on PC configs (motherboar,cpu,ram, net/audio/video devices, etc...,) but i don't see much why (if I understood well).

Webunny
Posts: 1201
Joined: Sat Apr 28, 2012 1:30 pm

Re: ROS testing on real hardware

Post by Webunny »

mametoc wrote:Well you know about "by people being interested in it" is not a suffice reason (like the mascot thing) but maybe someone can explain "ok, let's do it+quite reasons" or "directly not+few reasons".
The mascot thing has to be officially approved, because it IS an official thing; you link ROS and the mascot in an official/legal way, after all. That has nothing to do with making a wikipage about rigs that worked with ROS. I don't see why one would object to making a wikipage, if people want to create and maintain it, though. That's what a wiki is for, after all.

That said, I don't want to jump in head over toes; it's exactly that which z98 (and myself, btw) aren't all too fond off. For instance, I think a standard way (a form or something 'structured') would be needed, or at least is best suited for people wanting to share the knowledge. I therefore think it should encompass a structured way of indicating the what and who. I'm thinking something like this:

1)build/release number
2)model of computer, cpu (maybe GHZ and RAM, but that doesn't seem really all that relevant; maybe as option?)
3)motherboard
4)graphic card (or integrated graphics)
5)soundcard (or integrated)
6)extra comments


Something like that. I know it's currently only for a niche (most here use VM, but as proven by my testing, real HW tests divulge regressions that VM doesn't), but if I'm not mistaken, there are a few dedicated real HW testers among us as well. In the end, it will HAVE to work on real HW, so sooner or later it needs testing in the real world instead of a virtual one.
Last edited by Webunny on Thu Dec 19, 2013 10:00 pm, edited 2 times in total.

Z98
Release Engineer
Posts: 3379
Joined: Tue May 02, 2006 8:16 pm
Contact:

Re: ROS testing on real hardware

Post by Z98 »

There's a couple of matters to consider when trying to set up such things. First is the granularity of data, whether it is more useful to split compatibility on a per component basis or a system basis. This distinction is getting trickier and trickier with motherboard vendors jamming more stuff onto their boards. Another is how easy it is to do analysis on the data. A wiki page may seem the most accessible, but it's hard to scrape data off a wiki page and then do data analysis over it. What would be nice is to have a proper database backed repository, but developing one is kind of a full time web development project. Manpower, as always, is the critical weakness.

Webunny
Posts: 1201
Joined: Sat Apr 28, 2012 1:30 pm

Re: ROS testing on real hardware

Post by Webunny »

Z98 wrote:There's a couple of matters to consider when trying to set up such things. First is the granularity of data, whether it is more useful to split compatibility on a per component basis or a system basis. This distinction is getting trickier and trickier with motherboard vendors jamming more stuff onto their boards. Another is how easy it is to do analysis on the data. A wiki page may seem the most accessible, but it's hard to scrape data off a wiki page and then do data analysis over it. What would be nice is to have a proper database backed repository, but developing one is kind of a full time web development project. Manpower, as always, is the critical weakness.

Good points. That said, and as you indicate; such a more elaborate system might take a (very) long time. So maybe as a temporary, intermediary way of amassing this info, a wikipage would do. I think I may be willing to convert it into database backed repository when it ultimately made and the time has come for it.

That is, if at least 3-4 other real HW testers want to provide info and help with it too. (It's no use if I only do it myself).
Last edited by Webunny on Fri Dec 20, 2013 9:56 pm, edited 1 time in total.

milon
Posts: 969
Joined: Sat Sep 05, 2009 9:26 pm

Re: ROS testing on real hardware

Post by milon »

I'll be testing ROS on my 2 pc's, and maybe more depending on USB booting. :)

Webunny
Posts: 1201
Joined: Sat Apr 28, 2012 1:30 pm

Re: ROS testing on real hardware

Post by Webunny »

milon wrote:I'll be testing ROS on my 2 pc's, and maybe more depending on USB booting. :)
I've made a wikipage for real HW testing of complete PC systems: https://www.reactos.org/wiki/PC_ROS_Rigs

You and all real hardware testers can feel free to add theirs too.

Post Reply

Who is online

Users browsing this forum: Baidu [Spider] and 2 guests