Last night half way into a 32 player team game it became unplayable due to lag, maybe because it was a bigger map. I have an overclocked i5 3570K with some of the game's graphic settings turned down.
+3 / -0
|
I don't think it's a configuration problem. With 32 players and an average of 400 units in the game I think it's normal, I often have strong lag too. Perhaps as I have mentioned several times it would be better to go down to 10 v 10
+0 / -0
|
That is a 11 years old processor! I have a 5 year old processor and don't remember any big lag (https://cpu.userbenchmark.com/Compare/Intel-Core-i7-8700-vs-Intel-Core-i5-3570K/3940vs1316). Maybe it is time for an upgrade?...
+1 / -0
|
"Maybe it is time for an upgrade?" It's getting there. Computer is 11 years old but I've upgraded the RAM and graphics card since. Trying to get as much use as I can before replacing it, especially since Zero-K is the most taxing game I play.
+0 / -0
|
-
the game is still being developed so things that take up computing resources trickle in
-
new maps tend to be huge in size and unoptimized
-
engine development is increasingly targeting modern hardware
-
your machine probably suffers some wear and tear
+5 / -0
|
I found switching to linux and using the opensource GPU helped a lot with lag. I get 4-5 more fps per match, which does make a big difference when you've got 32 player matches with fleas flying everywhere. There's also the very very bad idea of disabling spectre mitigations to see if that works
+0 / -0
|
Usually with any software every now and then the more updates, more code, makes it slower. So in like 10 years this game will run like a turd on current hardware.
+1 / -0
|
A lot of the slowdown occurs from chili (the UI framework zk uses). Chili hasn't been updated (afaik) since 2018 or 2016 and optimizing it is a huge task that has a lingering threat of being dropped when/if engine releases their weird web language based ui framework.
+1 / -0
|
I upgraded to something twice as fast in all areas, and it made no difference to my game lag experience ! Now I've got a 3.6Ghz Xeon 6 core 6 thread CPU with 64Gb DDR4 RAM and SSD drives with a GTX 1080 6Gb GPU.
+0 / -0
|
I blame BAR maps with their extra metal => more units. For a long time I was playing on a stock i5-4570 (paired with a very modest Radeon RX550) and it has in recent times become a problem, even the i9-10850K that my ZK comrade plays on has started to chop a bit in some nasty situations (i.e. big games with lots of action). That was an expensive chip when new. I haven't tried the newer AMD stuff, but the 12th/13th gen Intel stuff is solid in ZK even at much better price points. I have played some hours on a i5-12500H (laptop CPU) and it handles things just as good as the i9-10850K despite being "only" mid-range and packaged into the small confines of a laptop. The GPU doesn't seem to matter, so long as it halfway decent and you run at 1920x res. quote:
Now I've got a 3.6Ghz Xeon 6 core 6 thread CPU with 64Gb DDR4 RAM and SSD drives with a GTX 1080 6Gb GPU. |
Adding cores is meaningless probably already by 3 of them. Same with RAM, if your program fits in RAM, you're good. Its basically a true/false situation. You want a CPU with high turbo, a modern core design, and probably a big L3. I wonder if those "X3D" AMD parts are good at this. Also, those top-end Intel chips that turbo to 6ghz, that would be smooooooth.
+0 / -0
|
No clue what you exactly mean by "Xeon 6". Do you mean "Xeon E3-12xx v6 series" (https://en.wikipedia.org/wiki/Xeon) ? That does not have a 3.6Ghz version though (ref: https://en.wikipedia.org/wiki/Kaby_Lake)Xeon is not primarily a desktop processor, it is a server processor (optimized for total processor performance and power consumption rather than single core performance). ZK will benefit more for higher single core performance, so does not amaze me that a Xeon is not a good option.
+0 / -0
|
@i9-10850K the i9 does not seem like that a big improvement over mine (Single core only 16%), even if it is 3 years newer than mine (ref: https://cpu.userbenchmark.com/Compare/Intel-Core-i9-10850K-vs-Intel-Core-i7-8700/m1255865vs3940). On the other hand, I do play on Linux. Played some times in windows as well but never tried to compare one with the other. Also, maybe you are spoiled and call lag what I would not notice. I generally do not like anyhow battles with X starlights and Z detris so maybe it's just that I do not get to be in the same massive battles that you are in.
+0 / -0
|
quote:
Also, maybe you are spoiled and call lag what I would not notice. |
:D I don't play on the big hardware usually, so I do get the lag, right up to losing control over my units. Last night I was on a Linux laptop with a i7-7820HQ and some "sufficient" nVidia GPU, it can certainly fall behind the server on a big crowded BAR map. quote:
No clue what you exactly mean by "Xeon 6" |
I expect this is a reference to the older Xeons that did double duty has HEDT desktop gamer chips. You know those big ones with 3- and 4-channel RAM. For example a Xeon W3690 or Xeon E5-1650 almost fits the description (but I don't quite care enough to search farther in that direction). The trouble with those is that, in ZK, they have no particular advantage over a regular desktop chip of the same era.
+1 / -0
|
quote:
@i9-10850K the i9 does not seem like that a big improvement over mine (Single core only 16%), even if it is 3 years newer than mine |
As you perhaps know, that was the era (which is sort of repeating itself in 12th, 13th, 14th gen) where Intel re-issued the same dogfood for years on end. So the i9-10850K came with 10 cores but they were individually pretty similar to anything else in the past 4 years. They did however turbo to slightly over 5ghz and the L3 was 20M, which is more than nothing, as improvements go. It is clearly superior to the i7-7820HQ despite being essentially the same thing at +25% clockspeed. This is partly why I theorize that a large L3 matters in ZK.
+0 / -0
|
--- You want a CPU with high turbo, a modern core design, and probably a big L3. --- Mine is: https://www.intel.com/content/www/us/en/products/sku/92994/intel-xeon-processor-e51650-v4-15m-cache-3-60-ghz/specifications.htmlL3 is 15Mb Max Turbo is actually 3.8Ghz, only 4Ghz if run in single thread mode ! Quad channel RAM. I'm testing out the lowest graphic setting in the game, to see if that makes any difference compared to the default settings. ( I suspect it might ! )
+0 / -0
|
|
|
I notice server catch up appears a lot faster with the lowest graphic settings.
+0 / -0
|
some settings are hardcoded and zk doesnt recogognise for example /water 0 /advmapshaders 0 /advmodelshaders 0 and those are main culprits sometimes the tracks are on. i play on ghetto pc (2 psu, thin client mobo, core2 duo, 8gb ddr2, and on top of that on linux, with mouse and keyboard transmitted to linux box with barrier kvm :D ) so over couple years now zk got fatter / stopped respecting minium graphics settings / added some stuff to what it thinks is minimal gpu settings / removed bitmaped trees which on certain maps are pain.
+0 / -0
|
quote:
Last night I was on a Linux laptop with a i7-7820HQ and some "sufficient" nVidia GPU, it can certainly fall behind the server on a big crowded BAR map. |
Did not try BAR in months but from what I remember there were much more units and stuff around (projectiles, missiles, etc.). So I might also notice lag in BAR. quote:
appears a lot faster |
You could just use a timer and see how much faster it actually is. Based on the many optical illusion images around, each time I can avoid relying on a brain for a measurement I think it is the sensible thing to do ;). Seems comparable to what I have (not many benchmarks but: https://www.cpubenchmark.net/compare/Intel-Xeon-E5-1650-v4-vs-Intel-i7-7700K-vs-Intel-i7-8700K/2838vs2874vs3098). If you feel lag and I don't maybe I am lucky because I am less sensitive. Anyhow don't think SSD matters, ZK takes couple of gigabytes or RAM only and does not load anything after start (I have 32GB so everything should be comfortable in RAM, don't expect you have lower than 8GB so don't think it would matter)
+0 / -0
|
Its still single core performance that counds? ZK definitely runs smoother since I upgraded from Ryzen 1700 to 5950x. I also notice, on medium-big matches, when power settings are set to powersave when its not running smoothly. Changing it balanced settings helps in most situations. In one of the recent matches, escalating to 2-4 supers, lagged me out a few times for some seconds, but it did catch up again. (gpu is rtx 2070 but it seems only cpu gets hogged hard.) (running windows 10 atm.)
+0 / -0
|
