Re: [hlds_linux] Gross over usage of Syscalls.
The CPU usage that SRCDS manages to chug is borderline suicidal for the end result, not to mention the countless places in the engine where memory continues to leak like a sieve I agree with this, in general, but I don't think 'syscalls' is the issue here. You need to exclude everything but the main thread to get rid of the numerous helper/packet threads that spend their time waiting for a signal or polling a socket. The 'gettimeofday' is more an example of lack of effort at profiling/optimizing srcds than it is a majority of CPU usage. The only real issue, then, I see with this profile is the fsync() calls. Something tells me those are occuring on the main thread. That's bad. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] Gross over usage of Syscalls.
While I realize syscalls are not the whole picture, it's a nice easy start. Note, the above was taken with libfasttime.so, without this, gettimeofday would dominate my cpu. 3182 futex(0xf6e6e33c, FUTEX_WAKE_PRIVATE, 1) = 0 3182 futex(0xf6e6e358, FUTEX_WAIT_PRIVATE, 13044897, {0, 49981604} 3182 <... futex resumed> ) = -1 ETIMEDOUT (Connection timed out) 3182 futex(0xf6e6e33c, FUTEX_WAKE_PRIVATE, 1) = 0 While there are quite a few time calls in between there, this is the abridged version. EAGAIN only appeared with socket/network related calls. I should probably also note my load wasn't nearly as high. I'll verify it's the same case when SRCDS is maxed later today. On Wed, May 25, 2011 at 9:53 AM, Gary Stanley wrote: > At 01:32 AM 5/25/2011, Kyle Sanderson wrote: >> >> While this has been discussed a number of times (atrocious over usage >> of gettimeofday has been worked around). Can there be something done >> regarding the abuse of syscalls? I mean, something like this >> http://i.imgur.com/PdWTB.png isn't really that great. For instance, >> every 1 out of 4 futex syscalls fail due to timing out (shown in that >> picture). This is just expensive and downright silly. Not much can be >> done on our end without the source to SRCDS, which no doubt will not >> be released. The CPU usage that SRCDS manages to chug is borderline >> suicidal for the end result, not to mention the countless places in >> the engine where memory continues to leak like a sieve. This effects >> and hurts everyone. The 'serious' people who host servers cannot use >> Windows due to the lack of Symbols. > > Abuse of syscalls? No offense, but strace/ktrace/truss et al only show > syscalls, not flow of code! It measures the amount of syscalls and usage of > them in CPU time. I could write a program to loop main() and chew up 99% CPU > and no syscalls will be generated. > > Maybe you could post what is timing out? It's probably doing EAGAIN over and > over.. > > strace -Ff -s 9 -p whatever > > > > > > ___ > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] Gross over usage of Syscalls.
At 01:32 AM 5/25/2011, Kyle Sanderson wrote: While this has been discussed a number of times (atrocious over usage of gettimeofday has been worked around). Can there be something done regarding the abuse of syscalls? I mean, something like this http://i.imgur.com/PdWTB.png isn't really that great. For instance, every 1 out of 4 futex syscalls fail due to timing out (shown in that picture). This is just expensive and downright silly. Not much can be done on our end without the source to SRCDS, which no doubt will not be released. The CPU usage that SRCDS manages to chug is borderline suicidal for the end result, not to mention the countless places in the engine where memory continues to leak like a sieve. This effects and hurts everyone. The 'serious' people who host servers cannot use Windows due to the lack of Symbols. Abuse of syscalls? No offense, but strace/ktrace/truss et al only show syscalls, not flow of code! It measures the amount of syscalls and usage of them in CPU time. I could write a program to loop main() and chew up 99% CPU and no syscalls will be generated. Maybe you could post what is timing out? It's probably doing EAGAIN over and over.. strace -Ff -s 9 -p whatever ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] Gross over usage of Syscalls.
At 01:32 AM 5/25/2011, Kyle Sanderson wrote: While this has been discussed a number of times (atrocious over usage of gettimeofday has been worked around). Can there be something done regarding the abuse of syscalls? I mean, something like this http://i.imgur.com/PdWTB.png isn't really that great. For instance, every 1 out of 4 futex syscalls fail due to timing out (shown in that picture). This is just expensive and downright silly. Not much can be done on our end without the source to SRCDS, which no doubt will not be released. The CPU usage that SRCDS manages to chug is borderline suicidal for the end result, not to mention the countless places in the engine where memory continues to leak like a sieve. This effects and hurts everyone. The 'serious' people who host servers cannot use Windows due to the lack of Symbols. Abuse of syscalls? No offense, but strace/ktrace/truss et al only show syscalls, not flow of code! It measures the amount of syscalls and usage of them in CPU time. I could write a program to loop main() and chew up 99% CPU and no syscalls will be generated. Maybe you could post what is timing out? It's probably doing EAGAIN over and over.. strace -Ff -s 9 -p whatever ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] Gross over usage of Syscalls.
Actually, I think there's a fairly good chance that the people there that do constructive work have been skinned and made into hats. On Wed, May 25, 2011 at 6:14 AM, Emil Larsson wrote: > Hats are useless if you don't have a stable server to play on. ;) Jokingly > aside, it's not like Valve programmers gets tied up when it comes to > cosmetics. It's more likely the engine programmers are more distracted on > Valve's other projects than TF2 cosmeticals. > > On Wed, May 25, 2011 at 8:29 AM, Evan wrote: > > > Hat sales = $$ vs fixing server = -$$ > > Everyone knows Valve has the best business model. > > > > > > On Wed, May 25, 2011 at 1:42 AM, John Marbury > > wrote: > > > > > I'm sure they'd love to fix this but OH LOOK A HAT!!! > > > > > > > > > On Wed, May 25, 2011 at 1:32 AM, Kyle Sanderson > > > wrote: > > > > > > > While this has been discussed a number of times (atrocious over usage > > > > of gettimeofday has been worked around). Can there be something done > > > > regarding the abuse of syscalls? I mean, something like this > > > > http://i.imgur.com/PdWTB.png isn't really that great. For instance, > > > > every 1 out of 4 futex syscalls fail due to timing out (shown in that > > > > picture). This is just expensive and downright silly. Not much can be > > > > done on our end without the source to SRCDS, which no doubt will not > > > > be released. The CPU usage that SRCDS manages to chug is borderline > > > > suicidal for the end result, not to mention the countless places in > > > > the engine where memory continues to leak like a sieve. This effects > > > > and hurts everyone. The 'serious' people who host servers cannot use > > > > Windows due to the lack of Symbols. > > > > > > > > It's a stability / performance trade off. Either way, the end > consumer > > > > is always the loser. Can this please be addressed? > > > > Kyle. > > > > > > > > ___ > > > > To unsubscribe, edit your list preferences, or view the list > archives, > > > > please visit: > > > > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > > > > > > > ___ > > > To unsubscribe, edit your list preferences, or view the list archives, > > > please visit: > > > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > > > > > ___ > > To unsubscribe, edit your list preferences, or view the list archives, > > please visit: > > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > > > ___ > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] Gross over usage of Syscalls.
Hats are useless if you don't have a stable server to play on. ;) Jokingly aside, it's not like Valve programmers gets tied up when it comes to cosmetics. It's more likely the engine programmers are more distracted on Valve's other projects than TF2 cosmeticals. On Wed, May 25, 2011 at 8:29 AM, Evan wrote: > Hat sales = $$ vs fixing server = -$$ > Everyone knows Valve has the best business model. > > > On Wed, May 25, 2011 at 1:42 AM, John Marbury > wrote: > > > I'm sure they'd love to fix this but OH LOOK A HAT!!! > > > > > > On Wed, May 25, 2011 at 1:32 AM, Kyle Sanderson > > wrote: > > > > > While this has been discussed a number of times (atrocious over usage > > > of gettimeofday has been worked around). Can there be something done > > > regarding the abuse of syscalls? I mean, something like this > > > http://i.imgur.com/PdWTB.png isn't really that great. For instance, > > > every 1 out of 4 futex syscalls fail due to timing out (shown in that > > > picture). This is just expensive and downright silly. Not much can be > > > done on our end without the source to SRCDS, which no doubt will not > > > be released. The CPU usage that SRCDS manages to chug is borderline > > > suicidal for the end result, not to mention the countless places in > > > the engine where memory continues to leak like a sieve. This effects > > > and hurts everyone. The 'serious' people who host servers cannot use > > > Windows due to the lack of Symbols. > > > > > > It's a stability / performance trade off. Either way, the end consumer > > > is always the loser. Can this please be addressed? > > > Kyle. > > > > > > ___ > > > To unsubscribe, edit your list preferences, or view the list archives, > > > please visit: > > > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > > > > > ___ > > To unsubscribe, edit your list preferences, or view the list archives, > > please visit: > > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > > > ___ > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] Gross over usage of Syscalls.
Stop jumping to conclusions. Email an engine programmer (Mike Durand, Zoid?) and they'll reply. They'll either show you how that's the best method or they know about the problem and are fixing it. On Wednesday, 25 May 2011, Evan wrote: > Hat sales = $$ vs fixing server = -$$ > Everyone knows Valve has the best business model. > > > On Wed, May 25, 2011 at 1:42 AM, John Marbury wrote: > >> I'm sure they'd love to fix this but OH LOOK A HAT!!! >> >> >> On Wed, May 25, 2011 at 1:32 AM, Kyle Sanderson >> wrote: >> >> > While this has been discussed a number of times (atrocious over usage >> > of gettimeofday has been worked around). Can there be something done >> > regarding the abuse of syscalls? I mean, something like this >> > http://i.imgur.com/PdWTB.png isn't really that great. For instance, >> > every 1 out of 4 futex syscalls fail due to timing out (shown in that >> > picture). This is just expensive and downright silly. Not much can be >> > done on our end without the source to SRCDS, which no doubt will not >> > be released. The CPU usage that SRCDS manages to chug is borderline >> > suicidal for the end result, not to mention the countless places in >> > the engine where memory continues to leak like a sieve. This effects >> > and hurts everyone. The 'serious' people who host servers cannot use >> > Windows due to the lack of Symbols. >> > >> > It's a stability / performance trade off. Either way, the end consumer >> > is always the loser. Can this please be addressed? >> > Kyle. >> > >> > ___ >> > To unsubscribe, edit your list preferences, or view the list archives, >> > please visit: >> > http://list.valvesoftware.com/mailman/listinfo/hlds_linux >> > >> ___ >> To unsubscribe, edit your list preferences, or view the list archives, >> please visit: >> http://list.valvesoftware.com/mailman/listinfo/hlds_linux >> > ___ > To unsubscribe, edit your list preferences, or view the list archives, please > visit: > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > -- Thanks, - Saul ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] Gross over usage of Syscalls.
Hat sales = $$ vs fixing server = -$$ Everyone knows Valve has the best business model. On Wed, May 25, 2011 at 1:42 AM, John Marbury wrote: > I'm sure they'd love to fix this but OH LOOK A HAT!!! > > > On Wed, May 25, 2011 at 1:32 AM, Kyle Sanderson > wrote: > > > While this has been discussed a number of times (atrocious over usage > > of gettimeofday has been worked around). Can there be something done > > regarding the abuse of syscalls? I mean, something like this > > http://i.imgur.com/PdWTB.png isn't really that great. For instance, > > every 1 out of 4 futex syscalls fail due to timing out (shown in that > > picture). This is just expensive and downright silly. Not much can be > > done on our end without the source to SRCDS, which no doubt will not > > be released. The CPU usage that SRCDS manages to chug is borderline > > suicidal for the end result, not to mention the countless places in > > the engine where memory continues to leak like a sieve. This effects > > and hurts everyone. The 'serious' people who host servers cannot use > > Windows due to the lack of Symbols. > > > > It's a stability / performance trade off. Either way, the end consumer > > is always the loser. Can this please be addressed? > > Kyle. > > > > ___ > > To unsubscribe, edit your list preferences, or view the list archives, > > please visit: > > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > > > ___ > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] Gross over usage of Syscalls.
I'm sure they'd love to fix this but OH LOOK A HAT!!! On Wed, May 25, 2011 at 1:32 AM, Kyle Sanderson wrote: > While this has been discussed a number of times (atrocious over usage > of gettimeofday has been worked around). Can there be something done > regarding the abuse of syscalls? I mean, something like this > http://i.imgur.com/PdWTB.png isn't really that great. For instance, > every 1 out of 4 futex syscalls fail due to timing out (shown in that > picture). This is just expensive and downright silly. Not much can be > done on our end without the source to SRCDS, which no doubt will not > be released. The CPU usage that SRCDS manages to chug is borderline > suicidal for the end result, not to mention the countless places in > the engine where memory continues to leak like a sieve. This effects > and hurts everyone. The 'serious' people who host servers cannot use > Windows due to the lack of Symbols. > > It's a stability / performance trade off. Either way, the end consumer > is always the loser. Can this please be addressed? > Kyle. > > ___ > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
[hlds_linux] Gross over usage of Syscalls.
While this has been discussed a number of times (atrocious over usage of gettimeofday has been worked around). Can there be something done regarding the abuse of syscalls? I mean, something like this http://i.imgur.com/PdWTB.png isn't really that great. For instance, every 1 out of 4 futex syscalls fail due to timing out (shown in that picture). This is just expensive and downright silly. Not much can be done on our end without the source to SRCDS, which no doubt will not be released. The CPU usage that SRCDS manages to chug is borderline suicidal for the end result, not to mention the countless places in the engine where memory continues to leak like a sieve. This effects and hurts everyone. The 'serious' people who host servers cannot use Windows due to the lack of Symbols. It's a stability / performance trade off. Either way, the end consumer is always the loser. Can this please be addressed? Kyle. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux