|
United Kingdom20345 Posts
"The following Gameplay options have been removed and are now permanently enabled"
I hate reading sentences like that
Fixed a performance issue relating to CPU thread balancing on newer CPU's that lessened performance over time
I wonder if that's live on beta yet - i've noticed a significant performance degradation and some inconsistency when trying to benchmark in the last few weeks! I re-run the same part of the replay as always, get the same results within 0.1% for 3 tests in a row but then the same thing a little later can give 15% different FPS. That didn't happen (as least as much) on earlier versions of the game
|
No water support for the editor yet? 
|
United Kingdom20345 Posts
Mining vespene gas from geysers of different orientations and locations should now be more consistent
Yay, no more having to put 4 workers on some gasses on a ton of maps in the game in order to get within 10% of the income that 3 are supposed to have
|
China6333 Posts
On October 06 2015 16:29 Cyro wrote:"The following Gameplay options have been removed and are now permanently enabled" I hate reading sentences like that Show nested quote +Fixed a performance issue relating to CPU thread balancing on newer CPU's that lessened performance over time I wonder if that's live on beta yet - i've noticed a significant performance degradation and some inconsistency when trying to benchmark in the last few weeks! I re-run the same part of the replay as always, get the same results within 0.1% for 3 tests in a row but then the same thing a little later can give 15% different FPS. That didn't happen (as least as much) on earlier versions of the game
Pretty sure patch 3.0 is a significantly newer build than beta. Keep in mind the beta build hasn't been updated since July 29th.
|
Czech Republic12134 Posts
On October 06 2015 15:17 Thaniri wrote:
I'd love for someone who is technologically competent to explain what the 64 bit change will mean.
Obviously now on our 64 bit processors the game should increase in performance, however with the engines well known problems with murdering CPUs could this increase of usage capability (wording??) make little to no difference?
The biggest advantage is that it(SC2) has more memory available from operating system. If you have 16 GB RAM then, in theory, the whole SC2 can be loaded into it(I hope I remember the SC2 size correctly, is it 9 GB, isn't it?  ). 32bit application has 1.75 GB limit of RAM usage(can be exceeded, but this type of compilation is not used that often).
There are other benefits too, but that's just nitpicking compared to this boost(from your view)
If it works properly( ) it should be faster and smoother. RAM is the fastest large memory place you have in PC.
(CPU cache is small(MB) and graphic memory is not generally accessible like RAM)
|
"Purchase of campaigne ingame" does this my friend who has no sc2 can buy lotv then buy the wol and hots campaigne for a bit less then the real games? Would be awesome since he is only interested in the campaignes.
|
•Many changes to the SC2 editor
This one is the most interesting
|
United Kingdom20345 Posts
On October 06 2015 17:01 digmouse wrote:Show nested quote +On October 06 2015 16:29 Cyro wrote:"The following Gameplay options have been removed and are now permanently enabled" I hate reading sentences like that Fixed a performance issue relating to CPU thread balancing on newer CPU's that lessened performance over time I wonder if that's live on beta yet - i've noticed a significant performance degradation and some inconsistency when trying to benchmark in the last few weeks! I re-run the same part of the replay as always, get the same results within 0.1% for 3 tests in a row but then the same thing a little later can give 15% different FPS. That didn't happen (as least as much) on earlier versions of the game Pretty sure patch 3.0 is a significantly newer build than beta. Keep in mind the beta build hasn't been updated since July 29th.
That's interesting
Also worthy of note, i'm getting ~6-7% higher performance on 32-bit than 64-bit at the moment on beta
|
It was an approx. 1.6GB download for me if that interests anyone out there on slow connections.
|
Canada8160 Posts
On October 06 2015 13:06 BigRedDog wrote:Too bad FFA is removed and i agree if no one plays, might as well get rid of it. I wonder how come no one plays FFA BGH $$$ maps no more or is it bc everyone gone to playing mono battles?? In WC3, i love FFA. SC2, not as much. Probably too focus on 1v1. Also, if Whisper of Oblivion is available for all players, won't those who pre-ordered LOTV early to play that will be upset? 
It was already stated that the missions would become free for everyone, whoever pre ordered the beta just got earlier access to them
|
Czech Republic12134 Posts
On October 06 2015 17:54 Cyro wrote:Show nested quote +On October 06 2015 17:01 digmouse wrote:On October 06 2015 16:29 Cyro wrote:"The following Gameplay options have been removed and are now permanently enabled" I hate reading sentences like that Fixed a performance issue relating to CPU thread balancing on newer CPU's that lessened performance over time I wonder if that's live on beta yet - i've noticed a significant performance degradation and some inconsistency when trying to benchmark in the last few weeks! I re-run the same part of the replay as always, get the same results within 0.1% for 3 tests in a row but then the same thing a little later can give 15% different FPS. That didn't happen (as least as much) on earlier versions of the game Pretty sure patch 3.0 is a significantly newer build than beta. Keep in mind the beta build hasn't been updated since July 29th. That's interesting Also worthy of note, i'm getting ~6-7% higher performance on 32-bit than 64-bit at the moment on beta
What is your measurement?
|
United Kingdom20345 Posts
On October 06 2015 18:20 deacon.frost wrote:Show nested quote +On October 06 2015 17:54 Cyro wrote:On October 06 2015 17:01 digmouse wrote:On October 06 2015 16:29 Cyro wrote:"The following Gameplay options have been removed and are now permanently enabled" I hate reading sentences like that Fixed a performance issue relating to CPU thread balancing on newer CPU's that lessened performance over time I wonder if that's live on beta yet - i've noticed a significant performance degradation and some inconsistency when trying to benchmark in the last few weeks! I re-run the same part of the replay as always, get the same results within 0.1% for 3 tests in a row but then the same thing a little later can give 15% different FPS. That didn't happen (as least as much) on earlier versions of the game Pretty sure patch 3.0 is a significantly newer build than beta. Keep in mind the beta build hasn't been updated since July 29th. That's interesting Also worthy of note, i'm getting ~6-7% higher performance on 32-bit than 64-bit at the moment on beta What is your measurement?
Just ran my benchmark on 32-bit and my performance was actually 10% higher than i've managed to record it before (while always benching in 64-bit), though my CPU is clocked 100mhz up and maybe has some other settings tweaked now.
tl;dr not a great scientific tests but obviously no huge performance improvements with 64
Ran both now to check without restarting system or changing other variables:
64 bit = 75 min, 106 average.
32 bit = 80 min, 113 average.
My regular benchmarks with a small range of settings were giving very exact results between about 65-72 min - so i've improved performance a little above my max on 64-bit already and then the 32 bit is significantly above that.
|
Czech Republic12134 Posts
On October 06 2015 18:31 Cyro wrote:Show nested quote +On October 06 2015 18:20 deacon.frost wrote:On October 06 2015 17:54 Cyro wrote:On October 06 2015 17:01 digmouse wrote:On October 06 2015 16:29 Cyro wrote:"The following Gameplay options have been removed and are now permanently enabled" I hate reading sentences like that Fixed a performance issue relating to CPU thread balancing on newer CPU's that lessened performance over time I wonder if that's live on beta yet - i've noticed a significant performance degradation and some inconsistency when trying to benchmark in the last few weeks! I re-run the same part of the replay as always, get the same results within 0.1% for 3 tests in a row but then the same thing a little later can give 15% different FPS. That didn't happen (as least as much) on earlier versions of the game Pretty sure patch 3.0 is a significantly newer build than beta. Keep in mind the beta build hasn't been updated since July 29th. That's interesting Also worthy of note, i'm getting ~6-7% higher performance on 32-bit than 64-bit at the moment on beta What is your measurement? Show nested quote +Just ran my benchmark on 32-bit and my performance was actually 10% higher than i've managed to record it before (while always benching in 64-bit), though my CPU is clocked 100mhz up and maybe has some other settings tweaked now.
tl;dr not a great scientific tests but obviously no huge performance improvements with 64
Ran both now to check without restarting system or changing other variables:
64 bit = 75 min, 106 average.
32 bit = 80 min, 113 average. My regular benchmarks with a small range of settings were giving very exact results between about 65-72 min - so i've improved performance a little above my max on 64-bit already and then the 32 bit is significantly above that.
So basically FPS only? Because I would expect lower FPS on 64bit but faster loading times and smoother play. I honestly don't care that much, I have 140+ FPS and SC2 is on Samsung 840Pro, so it is fast and furious But I hate waiting on other people loading so if that gets better for them with 64bit I would be thrilled
Nothing against you, I know that loading tests are annoying and smooth play testing is highly subjective, I was just curious, not nitpicking or anything
Edit> Or is that a point value?
|
United Kingdom20345 Posts
On October 06 2015 18:39 deacon.frost wrote:Show nested quote +On October 06 2015 18:31 Cyro wrote:On October 06 2015 18:20 deacon.frost wrote:On October 06 2015 17:54 Cyro wrote:On October 06 2015 17:01 digmouse wrote:On October 06 2015 16:29 Cyro wrote:"The following Gameplay options have been removed and are now permanently enabled" I hate reading sentences like that Fixed a performance issue relating to CPU thread balancing on newer CPU's that lessened performance over time I wonder if that's live on beta yet - i've noticed a significant performance degradation and some inconsistency when trying to benchmark in the last few weeks! I re-run the same part of the replay as always, get the same results within 0.1% for 3 tests in a row but then the same thing a little later can give 15% different FPS. That didn't happen (as least as much) on earlier versions of the game Pretty sure patch 3.0 is a significantly newer build than beta. Keep in mind the beta build hasn't been updated since July 29th. That's interesting Also worthy of note, i'm getting ~6-7% higher performance on 32-bit than 64-bit at the moment on beta What is your measurement? Just ran my benchmark on 32-bit and my performance was actually 10% higher than i've managed to record it before (while always benching in 64-bit), though my CPU is clocked 100mhz up and maybe has some other settings tweaked now.
tl;dr not a great scientific tests but obviously no huge performance improvements with 64
Ran both now to check without restarting system or changing other variables:
64 bit = 75 min, 106 average.
32 bit = 80 min, 113 average. My regular benchmarks with a small range of settings were giving very exact results between about 65-72 min - so i've improved performance a little above my max on 64-bit already and then the 32 bit is significantly above that. So basically FPS only? Because I would expect lower FPS on 64bit but faster loading times and smoother play. I honestly don't care that much, I have 140+ FPS and SC2 is on Samsung 840Pro, so it is fast and furious But I hate waiting on other people loading so if that gets better for them with 64bit I would be thrilled Nothing against you, I know that loading tests are annoying and smooth play testing is highly subjective, I was just curious, not nitpicking or anything
Loading felt the same, i didn't feel need to benchmark it considering it takes a couple seconds (would be very hard to get an exact result without a high speed camera) and just loading 1 replay doesn't hit RAM limits for 32 bit. I can't see why it would really be different and if it is, it's not significant.
Smoothness of play is absolutely not subjective, it's quite easily measurable by checking the frametime of every frame. I've done that quite a lot for sc2.
I have 140+ FPS
You don't maintain it and it's easy to tell the difference between 80fps and 120fps in sc2 even on a 60hz monitor because the frametimes are extremely uneven
|
Czech Republic12134 Posts
On October 06 2015 18:44 Cyro wrote:Show nested quote +On October 06 2015 18:39 deacon.frost wrote:On October 06 2015 18:31 Cyro wrote:On October 06 2015 18:20 deacon.frost wrote:On October 06 2015 17:54 Cyro wrote:On October 06 2015 17:01 digmouse wrote:On October 06 2015 16:29 Cyro wrote:"The following Gameplay options have been removed and are now permanently enabled" I hate reading sentences like that Fixed a performance issue relating to CPU thread balancing on newer CPU's that lessened performance over time I wonder if that's live on beta yet - i've noticed a significant performance degradation and some inconsistency when trying to benchmark in the last few weeks! I re-run the same part of the replay as always, get the same results within 0.1% for 3 tests in a row but then the same thing a little later can give 15% different FPS. That didn't happen (as least as much) on earlier versions of the game Pretty sure patch 3.0 is a significantly newer build than beta. Keep in mind the beta build hasn't been updated since July 29th. That's interesting Also worthy of note, i'm getting ~6-7% higher performance on 32-bit than 64-bit at the moment on beta What is your measurement? Just ran my benchmark on 32-bit and my performance was actually 10% higher than i've managed to record it before (while always benching in 64-bit), though my CPU is clocked 100mhz up and maybe has some other settings tweaked now.
tl;dr not a great scientific tests but obviously no huge performance improvements with 64
Ran both now to check without restarting system or changing other variables:
64 bit = 75 min, 106 average.
32 bit = 80 min, 113 average. My regular benchmarks with a small range of settings were giving very exact results between about 65-72 min - so i've improved performance a little above my max on 64-bit already and then the 32 bit is significantly above that. So basically FPS only? Because I would expect lower FPS on 64bit but faster loading times and smoother play. I honestly don't care that much, I have 140+ FPS and SC2 is on Samsung 840Pro, so it is fast and furious But I hate waiting on other people loading so if that gets better for them with 64bit I would be thrilled Nothing against you, I know that loading tests are annoying and smooth play testing is highly subjective, I was just curious, not nitpicking or anything Loading felt the same, i didn't feel need to benchmark it considering it takes a couple seconds (would be very hard to get an exact result without a high speed camera) and just loading 1 replay doesn't hit RAM limits for 32 bit. I can't see why it would really be different and if it is, it's not significant. Smoothness of play is absolutely not subjective, it's quite easily measurable by checking the frametime of every frame. I've done that quite a lot for sc2. You don't maintain it and it's easy to tell the difference between 80fps and 120fps in sc2 even on a 60hz monitor because the frametimes are extremely uneven
Well I don't know what are my frames during the game, it's not lagging and it feels OK all the time, so I don't care
Hmm, you're right about the smoothness.
Loading - that's the problem of our high tech machines, we don't see the difference Difference in 0.3 s can be 4 s for someone with worse PC.
(it is similar with our DB performance tuning at work, our general working DB has so few records that lowering by 10 reads can be 250,000 reads on customers side >< I spent 4 years arguing we need a bigger RD DB and we still don't have it)
Thanks for details!
|
China6333 Posts
On October 06 2015 17:54 Cyro wrote:Show nested quote +On October 06 2015 17:01 digmouse wrote:On October 06 2015 16:29 Cyro wrote:"The following Gameplay options have been removed and are now permanently enabled" I hate reading sentences like that Fixed a performance issue relating to CPU thread balancing on newer CPU's that lessened performance over time I wonder if that's live on beta yet - i've noticed a significant performance degradation and some inconsistency when trying to benchmark in the last few weeks! I re-run the same part of the replay as always, get the same results within 0.1% for 3 tests in a row but then the same thing a little later can give 15% different FPS. That didn't happen (as least as much) on earlier versions of the game Pretty sure patch 3.0 is a significantly newer build than beta. Keep in mind the beta build hasn't been updated since July 29th. That's interesting Also worthy of note, i'm getting ~6-7% higher performance on 32-bit than 64-bit at the moment on beta
I have access to a newer than the beta build of the game (but older than live patch 3.0) and can confirm performance is vastly improved on 64bit client.
|
United Kingdom20345 Posts
On October 06 2015 18:48 deacon.frost wrote:Show nested quote +On October 06 2015 18:44 Cyro wrote:On October 06 2015 18:39 deacon.frost wrote:On October 06 2015 18:31 Cyro wrote:On October 06 2015 18:20 deacon.frost wrote:On October 06 2015 17:54 Cyro wrote:On October 06 2015 17:01 digmouse wrote:On October 06 2015 16:29 Cyro wrote:"The following Gameplay options have been removed and are now permanently enabled" I hate reading sentences like that Fixed a performance issue relating to CPU thread balancing on newer CPU's that lessened performance over time I wonder if that's live on beta yet - i've noticed a significant performance degradation and some inconsistency when trying to benchmark in the last few weeks! I re-run the same part of the replay as always, get the same results within 0.1% for 3 tests in a row but then the same thing a little later can give 15% different FPS. That didn't happen (as least as much) on earlier versions of the game Pretty sure patch 3.0 is a significantly newer build than beta. Keep in mind the beta build hasn't been updated since July 29th. That's interesting Also worthy of note, i'm getting ~6-7% higher performance on 32-bit than 64-bit at the moment on beta What is your measurement? Just ran my benchmark on 32-bit and my performance was actually 10% higher than i've managed to record it before (while always benching in 64-bit), though my CPU is clocked 100mhz up and maybe has some other settings tweaked now.
tl;dr not a great scientific tests but obviously no huge performance improvements with 64
Ran both now to check without restarting system or changing other variables:
64 bit = 75 min, 106 average.
32 bit = 80 min, 113 average. My regular benchmarks with a small range of settings were giving very exact results between about 65-72 min - so i've improved performance a little above my max on 64-bit already and then the 32 bit is significantly above that. So basically FPS only? Because I would expect lower FPS on 64bit but faster loading times and smoother play. I honestly don't care that much, I have 140+ FPS and SC2 is on Samsung 840Pro, so it is fast and furious But I hate waiting on other people loading so if that gets better for them with 64bit I would be thrilled Nothing against you, I know that loading tests are annoying and smooth play testing is highly subjective, I was just curious, not nitpicking or anything Loading felt the same, i didn't feel need to benchmark it considering it takes a couple seconds (would be very hard to get an exact result without a high speed camera) and just loading 1 replay doesn't hit RAM limits for 32 bit. I can't see why it would really be different and if it is, it's not significant. Smoothness of play is absolutely not subjective, it's quite easily measurable by checking the frametime of every frame. I've done that quite a lot for sc2. I have 140+ FPS You don't maintain it and it's easy to tell the difference between 80fps and 120fps in sc2 even on a 60hz monitor because the frametimes are extremely uneven Well I don't know what are my frames during the game, it's not lagging and it feels OK all the time, so I don't care Hmm, you're right about the smoothness. Loading - that's the problem of our high tech machines, we don't see the difference Difference in 0.3 s can be 4 s for someone with worse PC. (it is similar with our DB performance tuning at work, our general working DB has so few records that lowering by 10 reads can be 250,000 reads on customers side >< I spent 4 years arguing we need a bigger RD DB and we still don't have it) Thanks for details!
![[image loading]](http://i.imgur.com/tTaO3KO.png)
Green = 32 bit, blue = 64 bit.
My start time on these 2 benchmarks was a tiny bit off so i moved one result ~10 pixels to match them up more. The other benchmarks showed the same performance increase and you can see that green is clearly higher (especially on the slow frames) so that doesn't change my confidence in the results. The gap between the fast and the slow frames is notably smaller on 32-bit as well - looks like it spends less time whatever it's doing with the CPU on the game tick frames.
Well I don't know what are my frames during the game, it's not lagging and it feels OK all the time, so I don't care
You would almost certainly care if you knew what "better" looked and felt like!
As you can see from looking at the pic, this is literally an "80fps" minimum for the green - the FPS display never drops below 80 - yet a TON of frames come in at 1/40'th to 1/60'th of a second. 80 min, 113 average and it's still very noticably unsmooth on 60hz. That's not even a maxed battle; it's an engagement with carriers in the midgame of a 1v1 map.
I didn't even notice some performance improvements myself, but when making a 40% performance upgrade there were so many "wow this is way smoother than i expected" moments. Two of the major ones were games with a lot of zerglings and flying a mutalisk flock around missile turrets but it affects the majority of games
|
Czech Republic12134 Posts
On October 06 2015 18:53 Cyro wrote:Show nested quote +On October 06 2015 18:48 deacon.frost wrote:On October 06 2015 18:44 Cyro wrote:On October 06 2015 18:39 deacon.frost wrote:On October 06 2015 18:31 Cyro wrote:On October 06 2015 18:20 deacon.frost wrote:On October 06 2015 17:54 Cyro wrote:On October 06 2015 17:01 digmouse wrote:On October 06 2015 16:29 Cyro wrote:"The following Gameplay options have been removed and are now permanently enabled" I hate reading sentences like that Fixed a performance issue relating to CPU thread balancing on newer CPU's that lessened performance over time I wonder if that's live on beta yet - i've noticed a significant performance degradation and some inconsistency when trying to benchmark in the last few weeks! I re-run the same part of the replay as always, get the same results within 0.1% for 3 tests in a row but then the same thing a little later can give 15% different FPS. That didn't happen (as least as much) on earlier versions of the game Pretty sure patch 3.0 is a significantly newer build than beta. Keep in mind the beta build hasn't been updated since July 29th. That's interesting Also worthy of note, i'm getting ~6-7% higher performance on 32-bit than 64-bit at the moment on beta What is your measurement? Just ran my benchmark on 32-bit and my performance was actually 10% higher than i've managed to record it before (while always benching in 64-bit), though my CPU is clocked 100mhz up and maybe has some other settings tweaked now.
tl;dr not a great scientific tests but obviously no huge performance improvements with 64
Ran both now to check without restarting system or changing other variables:
64 bit = 75 min, 106 average.
32 bit = 80 min, 113 average. My regular benchmarks with a small range of settings were giving very exact results between about 65-72 min - so i've improved performance a little above my max on 64-bit already and then the 32 bit is significantly above that. So basically FPS only? Because I would expect lower FPS on 64bit but faster loading times and smoother play. I honestly don't care that much, I have 140+ FPS and SC2 is on Samsung 840Pro, so it is fast and furious But I hate waiting on other people loading so if that gets better for them with 64bit I would be thrilled Nothing against you, I know that loading tests are annoying and smooth play testing is highly subjective, I was just curious, not nitpicking or anything Loading felt the same, i didn't feel need to benchmark it considering it takes a couple seconds (would be very hard to get an exact result without a high speed camera) and just loading 1 replay doesn't hit RAM limits for 32 bit. I can't see why it would really be different and if it is, it's not significant. Smoothness of play is absolutely not subjective, it's quite easily measurable by checking the frametime of every frame. I've done that quite a lot for sc2. I have 140+ FPS You don't maintain it and it's easy to tell the difference between 80fps and 120fps in sc2 even on a 60hz monitor because the frametimes are extremely uneven Well I don't know what are my frames during the game, it's not lagging and it feels OK all the time, so I don't care Hmm, you're right about the smoothness. Loading - that's the problem of our high tech machines, we don't see the difference Difference in 0.3 s can be 4 s for someone with worse PC. (it is similar with our DB performance tuning at work, our general working DB has so few records that lowering by 10 reads can be 250,000 reads on customers side >< I spent 4 years arguing we need a bigger RD DB and we still don't have it) Thanks for details! ![[image loading]](http://i.imgur.com/MOvBDvJ.png) Green = 32 bit, blue = 64 bit. My start time on these 2 benchmarks was a tiny bit off so i moved one result ~10 pixels to match them up more. The other benchmarks showed the same performance increase and you can see that green is clearly higher (especially on the slow frames) so that doesn't change my confidence in the results Show nested quote +Well I don't know what are my frames during the game, it's not lagging and it feels OK all the time, so I don't care You would almost certainly care if you knew what "better" looked and felt like!
I am pretty sure it is high enough for my lame playing
Edit: Also it is worth noting that I play protoss and my Zerg is somewhere on a silver level, so I just play the zerg as it is supposed to be. Just build a shitton of stuff and amove it! For the motherland!
(also known as the Zapp Brannigan tactic - I send wave after wave after wave...! )
|
China6333 Posts
The framepacing issue has been there since WoL beta and no improvements were made on that front I believe. All the engine improvements are put towards multithread optimization, better shading efficiency and maybe something minor I didn't notice.
|
What happened to the ladder revamp that we were supposed to be getting? I was sure it would come in the UI update?
|
|
|
|
|
|
EPT
