I officially hate W7

[Nothinman]

Try to remember that companies like MS that release an OS to the public as final are not going to release something that they know has problems .

Actually they do that all of the time. Every product ever produced has known issues when it goes RTM, just not anything that the development team, QA people or management consider show-stoppers.

That is where companies like Apple have it easy. They know exactly what hardware is in a specific computer and when they release the OS they know how it will perform.

In theory yes, but the new iPhone antenna thing seems to indicate otherwise. As does the myriad of overheating problems that the XBox continually has. At some point they decided that the risk of the problem affecting people was worth it and pushed it through anyway.

When I test software for Autodesk they provide me with a workstation already configured exactly like the ones the programmers use. That allows me to find bugs and flaws without it being a hardware or driver issue.

And that also means that you won't catch things related to other hardware or driver revisions. So while it may work just fine with the video card and drivers you tested it on the chances of every AutoCAD user out there having that exact same setup is just above 0.

That's why so many places, like MS, have public beta testing programs. To broaden their test cases and work out more of the uncommon bugs before the final release.

 

[Modelworks]

I did this and it is only showing events relating to W7 services/processes failing. Nothing explicitly relating to hardware.

That is where you start looking then. Usually event viewer will not show things like memory problems under hardware. Instead look under system for anything that isn't an information event.

If services or processes are failing then start reading up on what that process does and why it failed. Anything listed as an administrative event should be checked. I do it even on systems that are running fine to head off problems.

 

[Modelworks]

Actually they do that all of the time. Every product ever produced has known issues when it goes RTM, just not anything that the development team, QA people or management consider show-stoppers.

Issues like you are referring to are ones that occur on systems outside of the machines used internally. If QA is involved then the product is beyond the internal testing phases and is when pc software starts to show issues. The products may be shipped anyway but I have never seen an OS shipped where problems were known about systems internal to the developer unless it was very minor. A developer is not going to ship a product that does not work on their own systems.

In theory yes, but the new iPhone antenna thing seems to indicate otherwise. As does the myriad of overheating problems that the XBox continually has. At some point they decided that the risk of the problem affecting people was worth it and pushed it through anyway.

Some things don't show up in hardware testing. I did hardware testing for years and some of the things you would never expect to happen do. MS didn't know when they produced the final xbox for shipment the problems it had or they would have fixed it. It wasn't until the magnitude of the problem appeared that changes were made. No company wants to ship a faulty product. They don't just decide to risk it anyway if they intend to stay in business.

And that also means that you won't catch things related to other hardware or driver revisions. So while it may work just fine with the video card and drivers you tested it on the chances of every AutoCAD user out there having that exact same setup is just above 0.

It means that the programmers can develop the software based on systems they know are stable and provide the testers the same software to find bugs that will not be related to hardware a user may have. It is standard testing protocol for all the major vendors. Only after the software is working on those systems do they move it to outside testing. It is more work for the developers if everyone is using a different system when creating the program to draw a line . Instead unified hardware is used and when the program is stable on that it is moved to further testing. If they know the line drawing code is solid they don't have to go through it with future complaints.

That's why so many places, like MS, have public beta testing programs. To broaden their test cases and work out more of the uncommon bugs before the final release.

Betas show up problems that may not appear internally but have nothing to do with internal hardware test. Good examples are console development. They know the hardware and can develop for it knowing that what they produce is exactly what the user will receive. They still beta test the games because of game play issues, but because the hardware is unified they don't have to go back to that level for developing code.

The problem with MS OS is they don't specify the exact hardware the OS was qualified on down to things like ram , cpu, chipsets. They instead use WHQL which still has far too many variables to accurately reproduce what they use . The PC world will always be that way unless they start getting more detailed.

 

[Nothinman]

The products may be shipped anyway but I have never seen an OS shipped where problems were known about systems internal to the developer unless it was very minor. A developer is not going to ship a product that does not work on their own systems.

Of course they won't ship something that just plain doesn't work, but they always ship software with known issues and what they may consider "very minor" may be considered a show-stopper by someone else. The point was just that everything ships with known issues, software most of all because it's so simple to say "It'll be fixed in the next service pack.".

Betas show up problems that may not appear internally but have nothing to do with internal hardware test.

Maybe. There's always a chance that something works fine on your internal setup but breaks on every other setup in existence because of a bug in the particular version of your video cards firmware or drivers. You can't blindly rule anything out.

The problem with MS OS is they don't specify the exact hardware the OS was qualified on down to things like ram , cpu, chipsets. They instead use WHQL which still has far too many variables to accurately reproduce what they use . The PC world will always be that way unless they start getting more detailed.

Which is another reason to use Linux. 99% of the drivers are included with the kernel and maintained by the same people that maintain every other piece of the kernel as well. I really don't know how manufacturers manage to put out drivers as shitty as they do for Windows.

 

[Modelworks]

Which is another reason to use Linux. 99% of the drivers are included with the kernel and maintained by the same people that maintain every other piece of the kernel as well. I really don't know how manufacturers manage to put out drivers as shitty as they do for Windows.

Linux has the clear advantage in that area. With linux I can look up a driver and see explicitly what hardware is what designed for, chipsets, kernels and more.

 

[Nothinman]

Linux has the clear advantage in that area. With linux I can look up a driver and see explicitly what hardware is what designed for, chipsets, kernels and more.

And for 99% of your devices you don't need to worry about drivers at all, you just plug it in and it just works. The only ones that might consider special consideration are non-free drivers like those from nVidia and AMD/ATI. IMO the only area that Linux doesn't have a clear advantage is in some commercial application support.

 

[BarkingGhostar]

You don't say :hmm:

Now, I did try to run the memory per the spec on Wednesday night, but when things started going crzy on Thursday I was suspicious of the memory actually being capable of running at the stated spec when the motherboard seemed to think not.

That is where you start looking then. Usually event viewer will not show things like memory problems under hardware. Instead look under system for anything that isn't an information event.

If services or processes are failing then start reading up on what that process does and why it failed. Anything listed as an administrative event should be checked. I do it even on systems that are running fine to head off problems.

Sorry, I should have been more clear. I checked Event Manager for all critical events, including under System. What critical events were showing up were not pertaining to hardware.

Still, I'm glad the problem has been identified and the faltering hardware removed to gain a stable condition again. ^_^ Question remains how long should I let this go before moving forward with Office and Photoshop installs?

Anyone have a recommended freeware stress-testing tool I could use to test the overall computer?

 

[Steltek]

Anyone have a recommended freeware stress-testing tool I could use to test the overall computer?

Here is a site with all sorts of stability testing tools for overclockers, but they will also suffice for your needs:

http://files.extremeoverclocking.com/browse.php?c=18

As previously stated in the thread, MemTest86+ is great for testing the memory subsystem.

Stress Prime 2004 Orthos Edition (you want the SMP version to stress both cores of your i3) is good for testing CPU thermals - use the Blend test and it'll give your CPU and RAM a workout the likes of which it will probably never receive in real world useage. Be warned though, your CPU will get HOT - it is actually possible to damage a processor with this software if the heatsink isn't properly applied (as one of my friends found out with his brand new i7 a few months ago).

If your PC can get through a 9 continuous hour session of Stress Prime 2004 Orthos' Blend test with no errors, you ought to be more than good to go.

 

[sham63]

Been building since 1993 and you dont know how to properly diagnose hardware? WTF? When you build do just put *everything* together, install the OS, and call it a day? Shit, i POST the mobo with cpu and memory, then add hardware pieces, before i even think about mounting in the case. NO WAY have you been building that long...no way.

You may not believe this, but I have been buildinng since 1989, and I just put everything together in the case and install the OS. It seems to work for me. I haven't had very many problems doing it this way. But everyone has their own way of doing things, none of them are necessarily wrong.

 

[tommo123]

fair enough, i only started checking things on each build after one went wrong and wouldn't boot. was a PITA and was easier to take it apart and test on each step in the end.

still follow that process today.

 

[gpolubes]

Then why was the system stable for +4 weeks and only exhibited instabilities last night? In fact, the instabilities only came up after I started to install Adobe Reader.
.

w7 is the best MS OS since DOS 6.22

I have also many pretentions to it but all they are minor and adwantages are much stronger and signifficant.

Adobe Reader is one of the most buggy application in the w7 environment during monthes.

Each time when they suggest new update I believe that at least now they fixed their bugs and I am trying to install. Just few minutes ago I tried 9.33 version - in vain.

After each desperate trial I install again and again FoxitReader to throw out Adobe Reader plugins and to reassociate PDF files back to FoxitReader.

- Everything works fine.

Sincerely, --Gennadii

 

[Narse]

Now, I did try to run the memory per the spec on Wednesday night, but when things started going crzy on Thursday I was suspicious of the memory actually being capable of running at the stated spec when the motherboard seemed to think not.

Sorry, I should have been more clear. I checked Event Manager for all critical events, including under System. What critical events were showing up were not pertaining to hardware.

Still, I'm glad the problem has been identified and the faltering hardware removed to gain a stable condition again. ^_^ Question remains how long should I let this go before moving forward with Office and Photoshop installs?

Anyone have a recommended freeware stress-testing tool I could use to test the overall computer?

You should never ever rely on the event viewer to show you whats wrong. I have been working on corporate PCs and networks for over 15 years and maybe 3% of the time something in the event viewer would help, the rest of the time it is useless. Also your first post screams bad memory. But unless you had done the TS on this type of problem before I can see how it would be hard to see. But now you know and I bet you will start memory swapping as soon as you see this type of behavior!

 

[jtisgeek]

I build a new system first thing I do is try to break it with linpack and that's befor any overclocking have to know everythings good.

 

[Nothinman]

You should never ever rely on the event viewer to show you whats wrong. I have been working on corporate PCs and networks for over 15 years and maybe 3% of the time something in the event viewer would help, the rest of the time it is useless.

Sad but true. Despite the amount of eventlogs generated during normal use, most of them are pointless and don't tell you anything useful. One reason why fixing problems on Linux is usually so much simpler.

 

[RebateMonger]

It's not that the Event logs aren't useful, but they are seldom useful for the "really serious" errors that crash the entire PC. When the error involves memory or disk errors (which are often the problem with major crashes), then the PC freezes or reboots and there's no way for the OS to write the error to disk.

 

[Nothinman]

It's not that the Event logs aren't useful, but they are seldom useful for the "really serious" errors that crash the entire PC. When the error involves memory or disk errors (which are often the problem with major crashes), then the PC freezes or reboots and there's no way for the OS to write the error to disk.

I've been in this industry long enough to know that Windows' eventlogs are rarely useful. Sure they contain lots of data in some areas, but it's usually of the useless, "informational" kind. And usually the error entries are obtuse and cryptic, things like "the data is the error code" are just stupid.

Obviously, I would rather it sacrifice an eventlog entry for my data's safety, however Windows is already more liberal about writing to disk in bad circumstances than most other OSes.

 

[Fayd]

yep, lesson i learnt ages ago. never skimp on memory, or the psu.

i learned ages ago that memory from any manufacturer can go bad at any time.

buy what's cheap, fits necessary specs, and test it thoroughly.

 

[nastymoon]

The only real problem ive had with windows 7 is software compatibility otherwise I love the OS

 

[konakona]

And for 99% of your devices you don't need to worry about drivers at all, you just plug it in and it just works. The only ones that might consider special consideration are non-free drivers like those from nVidia and AMD/ATI. IMO the only area that Linux doesn't have a clear advantage is in some commercial application support.

heh, I wish I could have said that back when I gave up on getting my tv cards to work (not just one, two!) and no amount of fiddling with different versions of ndiswrapper / madwifi really allowed me to connect to my school's PEAP network (I must have tried like couple dozens of suggestions I have found in different forums, including here at no avail). Glad all that is over, when 9.10 came along that was FINALLY solved. Took them long enough, but at least it works now

Yeah, but aside from that I haven't really had any driver issues with linux. Except people nowadays don't usually have problems with windows drivers either unless it's some rushed video card drivers that aim to increase gaming performance. The state of gaming in linux is quite lacking no matter how you put it, so I think it is unfair to imply windows drivers are more problem-prone. You obviously have way more experience and knowledge than I could possibly imagine (I have never worked in the industry, whereas...), so I am not trying to argue with you on the technical level. That is just my humble observation from a layman's viewpoint

 

[Nothinman]

heh, I wish I could have said that back when I gave up on getting my tv cards to work (not just one, two!) and no amount of fiddling with different versions of ndiswrapper / madwifi really allowed me to connect to my school's PEAP network (I must have tried like couple dozens of suggestions I have found in different forums, including here at no avail). Glad all that is over, when 9.10 came along that was FINALLY solved. Took them long enough, but at least it works now

Yea, TV cards and some wifi cards fit into that last 1%. Any time you're looking at using ndiswrapper means something is wrong because you shouldn't be trying to use Windows drivers in Linux. It was a neat hack, but that's about it IMO.

Yeah, but aside from that I haven't really had any driver issues with linux. Except people nowadays don't usually have problems with windows drivers either unless it's some rushed video card drivers that aim to increase gaming performance. The state of gaming in linux is quite lacking no matter how you put it, so I think it is unfair to imply windows drivers are more problem-prone. You obviously have way more experience and knowledge than I could possibly imagine (I have never worked in the industry, whereas...), so I am not trying to argue with you on the technical level. That is just my humble observation from a layman's viewpoint

It's not unfair at all, I've been using Linux on my desktop at home for about a decade now and the only driver issues I've really had were with the non-free nVidia ones and even those are only occasional and usually caused by me updating to a brand new kernel that's incompatible with their last release. But even at work I'm constantly seeing people with network/wifi issues, printer driver problems, etc. The quality of drivers just seems so very much lower on Windows in general.

And yes, the state of gaming on Linux is kind of sad. WINE is pretty much 50/50, sometimes it works just fine and sometimes it won't work no matter how much you bang your head on the keyboard. But whenever I tried Windows at home it was a similar experience only with different components. Finding the right combination of drivers and software for some games on Windows is like a game of roulette in itself. And I'd rather fight with Linux and WINE than Windows and drivers. =)

 

[n7]

Good to hear you go it figured out.

But:
Learn to test your hardware!

Assuming everything is fine is precisely where you went wrong.
Not once do i see where you post that you ran hardware testing...

It's always the first thing to do when things don't work as they should.

Assuming the OS is to blame is honestly ridiculously naive, as 99.99% of the time, if an OS is configured/installed properly on FUNCTIONING hardware, it'll work just fine.

There are countless free applications out there to test components, some general ones, some manfacturer-specific, etc.

Examples:
CPU stability testing (for OCing, etc)
-Prime 95
-LinX/Intel Burn Test
-
Memory (IMPERATIVE TO TEST THIS WITH A NEW SETUP!)
Memtest86+
HCI Memtest (much better when you have a lot of RAM to test)

HDD
-most manufacturers have their own!
Hiren's has a bunch
-MHDD
-HDDScan
-Speedfan/CrystalDiskInfo for reading SMART

GPU
-RTHDRIBL
-Furmark
-loop intensive games, etc.

And there are many more.
In general though, learning how to test your own hardware is pretty much needed if you want to happily build your own systems, as stuff fails...& when it does, if you don't know how to diagnose what is wrong...well...

 

[konakona]

TV cards are kinda niche to begin with, so I totally understand lack of interest/motivation to get that working. But hey wireless NICs? I know I am kind of nitpicking here though, seeing how that had been brought up as the one of very few major flaws with ubuntu for mainstream use.

Let me put this in a slightly different tone - I haven't found anything wrong with linux drivers in my personal experience aside those little buggers mentioned above, neither have I found anything on the windows side. Not sure how much of those problems people were having network / printers could be traced back to driver issues? Maybe, maybe not?

There are quite a few games that have some issues with specific driver releases, but I tend to think of it more as a game specific bugs. Usually things start getting ugly when there is an experimental update with huge boost - there is always a tried and true version to fall back onto. Either that, or game itself needs to get patched up. On the whole I think nv/ati driver problems are a bit exaggerated, caused by human error (improper uninstallation/installation and such), etc etc.

From a developer's standpoint, linux is way more debug-friendly if I am reading your post right?

 

[Nothinman]

But hey wireless NICs? I know I am kind of nitpicking here though, seeing how that had been brought up as the one of very few major flaws with ubuntu for mainstream use.

From an enduser perspective, I understand completely. But the problem is that so many manufacturers "wireless cards" are really just radios with all of the real work being done in the driver or the firmware and then they take it an extra step and bundle the firmware inside of the driver. So every time the driver starts up it pushes the firmware onto the wireless card and when the card is powered off it loses it's firmware. And in some cases the firmware couldn't be legally redistributed or the redistribution license was prohibitively vague. Broadcom was especially bad in this respect, I will never buy anything from them. Thankfully, most of them have caught on to how dumb this is and have separate, redistributable firmware now but I'm sure there's still a few stragglers.

Not sure how much of those problems people were having network / printers could be traced back to driver issues? Maybe, maybe not?

We were told by HP tech support to remove all of their drivers and support software and start from scratch with their drivers in one specific problem that we were fighting with over the last week or so. Just the fact that drivers for the those crap all-in-one printers are several hundred megs in size should be all you need to know.

There are quite a few games that have some issues with specific driver releases, but I tend to think of it more as a game specific bugs.

Which I have yet to see in Linux. Every Linux game that I've played that used OpenGL and/or SDL has worked just fine. And ironically most of the games I've played in WINE work fine too, although that can change from one WINE version to the next. APIs like DirectX are supposed to make bugs like that less prevalent, not more.

On the whole I think nv/ati driver problems are a bit exaggerated, caused by human error (improper uninstallation/installation and such), etc etc.

In general you're probably right, but the last time I tried gaming on Windows there wasn't much room for user error, the installation only consisted of newly installed Windows and the game I was trying to play. I didn't install anything else because all of my other stuff is done in Linux.

From a developer's standpoint, linux is way more debug-friendly if I am reading your post right?

That's an enormous understatement IMO.