I forgot to mention running 2 Milkyway tasks concurrently with my setup reduces the load on the server a bit because it means 2 or 3 tasks are downloaded at a time instead of 1 or 2.

Maybe there should be some sort of FAQ or README that goes with the GPU applications to explain things like this. It may cut down on some of the angst.

Actually there is a readme included in every download of the GPU applications that explains these things. The people only have to read it ;)

Pfff read, omg he said read and look something up whats that xD
Serious Paul, CP stop saying such nasty things ;)

Well, you are not the first to point out that I am a mean evil person ...

You probably won't be the last ...

ANd yes, I saw the humor in what you said ... can you see it what I said? :)

I've just put app_info for MW to run only 1 wu, also changed it to be 1 ATI and unsuspended Collatz. The MW wu's were placed into waiting and 2 Collatz wu's were kicked off. No biggie, but it does mean the 2 MW wu's which were at 80% will start fresh, bit I digress. Once one Collatz completed, BOINC restarted a MW wu on the same GPU as Collatz was running on leaving the other GPU idle. Collatz then slowed to a crawl and the MW was taking more than twice as long as it should.

6.10.13 is running.

but if the pros of doing it are outweighed by the negative side-effects, then why put it in there?

Because people wanted it ... and it may increase throughput in some cases ... if I am running tasks off at 55 seconds or less then a one second improvement in hand-over from one MW task to another is pretty significant... for those that want maximum performance/throughput...

But like all settings it is use at your own risk as is all anon platform installs. If it does not work right it is on your head.

I would rather CP (Gispel) and UCB gave us more options than less because there are many situations where the available controls are just not good enough to let us manage the operation of the system. Sadly the UCB philosophy is "trust us, we be smarter than you" rather than adding some common sense controls that would allow us flexibility ...

Sometimes developers actually do know better than end-users / business users. In my former job, business users told us to do multiple things that we felt were not the best way to go, then we'd get negative feedback from the actual day-to-day end users who didn't like it, which we'd then have to say that we were just doing as we were instructed by their management chain... We'd also get demands from the end users that were also inefficient, but their management would back them up on it and we'd be made to make changes that actually slowed down the system...and then everyone would fuss about how slow the system had become...

From what I'm seeing from you and CP, I think the "feature" is probably doing more harm than good, and thus should really be a candidate for review as to exactly how beneficial it really is vs. how much support problems and other angst it is causing...

Also, BOINC is a "for the masses" thing, and generally speaking, "the masses" just want it to run with the minimum amount of changing settings. The "power users" might do well with that fork that was mentioned... With all the years of use, as for the BOINC client itself, I've never needed all the bells and whistles that you all want, or I've been blissfully ignorant of my need... Of course, I realize that I don't need Vista / 7 support, nor do I need native GPU support, so I'm still back on the "ancient" 5.8.16 version... :shrug:

Sometimes developers actually do know better than end-users / business users. In my former job, business users told us to do multiple things that we felt were not the best way to go, then we'd get negative feedback from the actual day-to-day end users who didn't like it, which we'd then have to say that we were just doing as we were instructed by their management chain... We'd also get demands from the end users that were also inefficient, but their management would back them up on it and we'd be made to make changes that actually slowed down the system...and then everyone would fuss about how slow the system had become...

I have my horror stories and tales of woe from when I was a software developer too ... and yes I recall many times the "need" to code into place very inefficient mechanisms ...

From what I'm seeing from you and CP, I think the "feature" is probably doing more harm than good, and thus should really be a candidate for review as to exactly how beneficial it really is vs. how much support problems and other angst it is causing...

Maybe, maybe not ... if the user is only running MW it could be a desireable feature for some users. That it causes problems in multi-project use is not a certain indicator for removal.

Also, BOINC is a "for the masses" thing, and generally speaking, "the masses" just want it to run with the minimum amount of changing settings. The "power users" might do well with that fork that was mentioned... With all the years of use, as for the BOINC client itself, I've never needed all the bells and whistles that you all want, or I've been blissfully ignorant of my need... Of course, I realize that I don't need Vista / 7 support, nor do I need native GPU support, so I'm still back on the "ancient" 5.8.16 version... :shrug:

Again I agree with caveats, because sometimes the features desired would allow better operation with less monitoring. As to 5.8.16, if it is working for you then I always suggest that you stick with what is working...

You may also need his latest beta .20b for tasks here

And where can that be found?

On the MW server ;)

But before you try to download it, it's not complete and not correctly set up yet. So it will probably not work right now. Be a little bit more patient and it will be automatically distributed by the project if you have one of the latest client versions. And one doesn't have to copy the dlls anymore ;)

If you can't wait, you find it also here. Be sure to pick the right version. Look to the readme for details!

If you can't wait, you find it also here. Be sure to pick the right version. Look to the readme for details!

Of course I can't wait, and thanks! Both my boxes showed a drop of 3 seconds/WU.

Question in general about the MW & Collatz apps. Since v2 Collatz has been more or less stable on my machines (both HD 4770 / WinXP32 / v9.8 no CCC / BOINC v6.10.13). MW runs fine when the machines are left as dedicated crunchers, but if I try to do anything else when MW is running (like open Windows explorer or running notepad, sometimes even bringing BOINC manager into focus) I'll get an immediate video crash. Given buggy ATI drivers, what could be the difference that allows Collatz to run with other apps and causes MW to crash? Possibly that MW is using double precision?

Question in general about the MW & Collatz apps. Since v2 Collatz has been more or less stable on my machines (both HD 4770 / WinXP32 / v9.8 no CCC / BOINC v6.10.13). MW runs fine when the machines are left as dedicated crunchers, but if I try to do anything else when MW is running (like open Windows explorer or running notepad, sometimes even bringing BOINC manager into focus) I'll get an immediate video crash. Given buggy ATI drivers, what could be the difference that allows Collatz to run with other apps and causes MW to crash? Possibly that MW is using double precision?

I have really no idea. Both applications access the GPU in exactly the same way. The only difference is what gets calculated. And I don't have those problems when using Catalyst 8.12. ATI apparently changed something in the drivers, what has broken them for MW on XP. Vista and Win7 drivers appear not to be affected and strangely it doesn't happen with Collatz either. I have a WinXP machine running Collatz with Catalyst 9.9 (and a 790GX chipset IGP needing just 11,000 seconds on average ;) without any problems.

Question in general about the MW & Collatz apps. Since v2 Collatz has been more or less stable on my machines (both HD 4770 / WinXP32 / v9.8 no CCC / BOINC v6.10.13). MW runs fine when the machines are left as dedicated crunchers, but if I try to do anything else when MW is running (like open Windows explorer or running notepad, sometimes even bringing BOINC manager into focus) I'll get an immediate video crash. Given buggy ATI drivers, what could be the difference that allows Collatz to run with other apps and causes MW to crash? Possibly that MW is using double precision?

I have really no idea. Both applications access the GPU in exactly the same way. The only difference is what gets calculated. And I don't have those problems when using Catalyst 8.12. ATI apparently changed something in the drivers, what has broken them for MW on XP. Vista and Win7 drivers appear not to be affected and strangely it doesn't happen with Collatz either. I have a WinXP machine running Collatz with Catalyst 9.9 (and a 790GX chipset IGP needing just 11,000 seconds on average ;) without any problems.

Unfortunately Catalyst 8.12 isn't an option for the HD 4770. Maybe I'll give Win 7 a try when I get some free time. As long as I don't touch the mouse or keyboard they'll run MW for days without a video reset. Also have a 790GX IGP running Collatz on my main machine with no problems, but it's taking about 50% longer than yours. Thanks for the info!

Also have a 790GX IGP running Collatz on my main machine with no problems, but it's taking about 50% longer than yours. Thanks for the info!

v2.05b and some 900MHz+ clocks ;). But it's memory bandwidth limited with the Athlon 64 X2 in the box (shared memory bandwidth over HT is just 4GB/s). As the IGP core easily overclocks to 1GHz (with a modest voltage increase) even 10,000 seconds should be possible with a Phenom or the newer K10 based Athlon II X2, X3 or X4 as HT3.0 really helps the available bandwidth to the IGP.

v2.05b and some 900MHz+ clocks ;). But it's memory bandwidth limited with the Athlon 64 X2 in the box (shared memory bandwidth over HT is just 4GB/s). As the IGP core easily overclocks to 1GHz (with a modest voltage increase) even 10,000 seconds should be possible with a Phenom or the newer K10 based Athlon II X2, X3 or X4 as HT3.0 really helps the available bandwidth to the IGP.

Does your board have much sideport memory? I would think that would make a huge difference. Mine has only 64MB sideport, with a Phenom 9600 BE. Maybe if I disable everything but the sideport, or maybe swap in some faster memory...

v2.05b and some 900MHz+ clocks ;). But it's memory bandwidth limited with the Athlon 64 X2 in the box (shared memory bandwidth over HT is just 4GB/s). As the IGP core easily overclocks to 1GHz (with a modest voltage increase) even 10,000 seconds should be possible with a Phenom or the newer K10 based Athlon II X2, X3 or X4 as HT3.0 really helps the available bandwidth to the IGP.

Does your board have much sideport memory? I would think that would make a huge difference. Mine has only 64MB sideport, with a Phenom 9600 BE. Maybe if I disable everything but the sideport, or maybe swap in some faster memory...

I have 128MB sideport memory running with just below 1 GHz (effective clock), it overclocks together with the reference clock on my board, so its 246*4 right now ;).

The sideport itself is actually quite slow (has just a 32bit or even only 16 bit wide interface). It normally holds the framebuffer. That has the advantage that the memory controller in the CPU has nothing to do when there is just some 2D stuff going on. And for 3D (or GPGPU) it provides a bit of additional bandwidth as the IGP can use both in parallel. But the link to the memory of the CPU has actually a higher bandwidth than the sideport memory. So it's most likely not that important. I would think running sideport only is going to be slower than UMA only (but never tested it). Sideport + UMA should be the optimal setting.

By the way, for 10,000 seconds a memory bandwith of about 4GB/s is needed.

Question in general about the MW & Collatz apps. Since v2 Collatz has been more or less stable on my machines (both HD 4770 / WinXP32 / v9.8 no CCC / BOINC v6.10.13). MW runs fine when the machines are left as dedicated crunchers, but if I try to do anything else when MW is running (like open Windows explorer or running notepad, sometimes even bringing BOINC manager into focus) I'll get an immediate video crash. Given buggy ATI drivers, what could be the difference that allows Collatz to run with other apps and causes MW to crash? Possibly that MW is using double precision?

I have really no idea. Both applications access the GPU in exactly the same way. The only difference is what gets calculated. And I don't have those problems when using Catalyst 8.12. ATI apparently changed something in the drivers, what has broken them for MW on XP. Vista and Win7 drivers appear not to be affected and strangely it doesn't happen with Collatz either. I have a WinXP machine running Collatz with Catalyst 9.9 (and a 790GX chipset IGP needing just 11,000 seconds on average ;) without any problems.

I'm not entirely convinced that the CAL bugs introduced in ATI drivers earlier this year are fully exterminated. After I switched from BOINC 6.4.7 to 6.10.13, I tried both Cat 8.12 and 9.9 for a while on a HD4870, everything else unchanged.

Cat 8.12 rock solid up to 790MHz.
Cat 9.9 barely stable at stock, VPU crash guaranteed anywhere beyond 750Mhz.

---

Question in general about the MW & Collatz apps. Since v2 Collatz has been more or less stable on my machines (both HD 4770 / WinXP32 / v9.8 no CCC / BOINC v6.10.13). MW runs fine when the machines are left as dedicated crunchers, but if I try to do anything else when MW is running (like open Windows explorer or running notepad, sometimes even bringing BOINC manager into focus) I'll get an immediate video crash. Given buggy ATI drivers, what could be the difference that allows Collatz to run with other apps and causes MW to crash? Possibly that MW is using double precision?

I have really no idea. Both applications access the GPU in exactly the same way. The only difference is what gets calculated. And I don't have those problems when using Catalyst 8.12. ATI apparently changed something in the drivers, what has broken them for MW on XP. Vista and Win7 drivers appear not to be affected and strangely it doesn't happen with Collatz either. I have a WinXP machine running Collatz with Catalyst 9.9 (and a 790GX chipset IGP needing just 11,000 seconds on average ;) without any problems.

I'm not entirely convinced that the CAL bugs introduced in ATI drivers earlier this year are fully exterminated. After I switched from BOINC 6.4.7 to 6.10.13, I tried both Cat 8.12 and 9.9 for a while on a HD4870, everything else unchanged.

Cat 8.12 rock solid up to 790MHz.
Cat 9.9 barely stable at stock, VPU crash guaranteed anywhere beyond 750Mhz.

830 GPU, 1050 Memory clocks on a 4870. CCC 9.9, BOINC 6.10.6. Rock solid for days now on 85% CPU (all 4 cores). I run one Milky Way wu at a time at an average of about 43 seconds. I like the 1 second pause of juice going to the chip to keep it at 76C at 85% fan...keeping room comfortable!

RAM running at 3.7 GB out of 8 at 800Mhz. Vista x64. Would love Linux version...I get 30% better CPU benchmarks in Ubuntu.

I get 30% better CPU benchmarks in Ubuntu.

And how does that help your GPU to calculate the WUs?

To say the truth, the meaningless BOINC benchmark value doesn't help the CPU projects neither ;). I would really like if Berkeley scraps it completely :]

CPDN has ignored it from day one, just as it has ignored deadlines.

---

Cheers,

PeterV

.

lol i agree with u CP the boinc bench is as usefull as a ati 9700xt for boinc.

As to the issues with the 4770/4890 somehow these cards seems to only work kinda stable on vista or win7 and in most cases just in the X64 versions.

I have helped many people with getting it to run on their system, but somehow the brand of the cards does make a difference.
Most Saphire/XFX cards give much less problems then other brands i have encountered.
Nevertheless when i get them running stable and they are not touched by windows update or driver updates most run without a hitch.

But some i can't find whats bugging them i have seen some club3d card 4770 which give constant vpu recover errors on almost everything i did on them.
But when you see how fast the screens build and how most games run on them you are baffled why they fail at collatz/mw.

I start to think some cards have some changes to the pcb which make it hard to get them to work right at D.C.

---

Its new, its relative fast... my new bicycle

To say the truth, the meaningless BOINC benchmark value doesn't help the CPU projects neither ;). I would really like if Berkeley scraps it completely :]

I'll drink to that any day. It would be much better to get rid of the whole broken BOINC benchmark system. Have the projects institute fixed or server assigned credit as so many do already.