Jump to content


WOT 1.3.0.1 vs WOT 1.4 CT with ACTIVE multi-threading [HUGE gain of FPS]


  • Please log in to reply
45 replies to this topic

ValkyrionX #1 Posted 16 January 2019 - 02:58 PM

    Captain

  • Player
  • 52677 battles
  • 2,462
  • [RDDT] RDDT
  • Member since:
    02-07-2015

Finally I found the time to do this thread, with the hope that it will be useful as a comparison to those who did not and will not have time to try the 1.4 patch with multi-threading support in the game engine.

We all appreciated the release of the patch 1.0 at least for what concerned the new game engine and the finally modern graphics of the game, without clearly forgetting the players with low-powered PCs that have always been under the attention of the WG and that even today can play thanks to the SD client with minimal settings and the aid of graphic mods to further reduce the graphic quality to increase performance.
If we want to give credit to the WG [sometimes we have to grant it] this is certainly a merit that can not be denied.

Will try both HD clients with ultra settings

 

-

Let's move on to the facts.

-

Here I will propose the comparison between the performance of the current game in version 1.3 compared to that of version 1.4-CT on the test server.
The multithreading support, as you will see later, has finally granted the CPU of my computer to be used properly, thus freeing the resources [so far bottled] of my graphics card.

 

_________________________________________________________________________________________________________________________________________________________________________

 

The following configuration of my [old] computer, used for both comparisons is as follows:

 

MOTHERBOARD

Asrock K10N78M-PRO

https://www.asrock.c...M Pro/index.asp

 

C.P.U

AMD Phenom II 965 Black Edition 3,4GHz [x4 Core] [x4 Threads] 125 Watt-64 bit-1,40 volts [AM3 processor]

https://www.amd.com/...roducts/cpu/965

 

R.A.M

Kingston HyperX Genesis DDR2 Kit of 2 banks 2GBx2 [tot. 4GB] 1066MHz - Dual Channel - timings 5-5-5-15-25 at 2,2 volts - mode "Address 6 bit" - Unganged - Unbuffered - NON ECC - CR2

https://www.kingston...00D2LLK2_4G.pdf

 

G.P.U

AMD Sapphire Radeon RX570 Nitro+ 4GB GDDR5 [GPU Clock 1340MHz - VRAM Clock 1750MHz] 135Watt - 256bit 

http://sapphirenitro...m/en/570-4.html

 

SSD

Kingston SSDNow300 120GB - AHCI mode- SATA 2.0

 

Power Supply

Corsair TX650 Watt [80 Plus Gold version]

https://www.corsair....p/CP-9020038-NA

 

Monitor

Philips 227EL LED

- 1920x1080 8bit - via DVI clabe - 65Hz [no gsync - no amd freesync]

 

SO

Windows 10 Pro 64bit 

 

_________________________________________________________________________________________________________________________________

*Note that I did not overclock the graphics card, but only the 3.4 Ghz to 3.7ghz CPU with 1,40 volts
Same thing goes for the motherboard, I brought the NB bus speed from 1800MHz to 2000MHz and that of the HT Link from 1800 always to 2000MHz

the ram is native for 800MHz dual channel, but the hyperX genesis model is set to work at 1066MHz with 2.2 voltage increase from 1.80v stock

these light overclockrs have greatly increased the communication speed between the CPU and the components on the mobo, such as the RAM and the GPU
I have reduced the latency of the ram from about 80ns to about 58ns and the response time and processing of the CPU compared to 3.4 ghz has slightly improved
the most significant increase was the latency gain, write / read speed and copy of the ram in communication with CPU and SSD

___________________________________________________________________________________________________________________________________

 

 

 

Some image to get an idea of how it is physically my computer:

 

*I apologize for the bad quality of the pics, my girlfriend is playing WOT with her laptop next to my pc and I do not have a proper camera

 

Pics

 

 

 

 

WOT ENCORE SCORE WITH MY HARDWARE

 

 

 

 

 

 

WOT 1.3 vs WOT 1.4

 

 

 

Memory consumption [RAM and VRAM]

 

Verdict:

Almost the same

 

 

1.3

 

 

1.4

 

 

 

 

 

CPU and GPU usage during the game

 

Verdict:

 

The version 1.4 reduces the spasmodic use of the processor core 0, redistributing the work on all available cores in a more balanced way.

certainly the core 0 still suffers from a higher workload than the 1-2-3 cores, but compared to version 1.3 the core 0 is no longer fixed at 100% of workload leaving the remaining cores unused, or almost. 

surely over time with subsequent patches we can get even better results
This subsequently affects the work of calculating the GPU which is reduced in its total load, thus increasing, compared to version 1.3, the resources to be able to devote more sensibly to 3D in the version without multithreading.

In other words, this frees up more resources of our GPU thanks to a balanced use of the central processor, clearly only if multicores

forget about superlative improvements with a processor that does not have at least 3 or 4 cores

 

 

 

1.4

 

 

1.3

 

 

 

 

FPS

 

Verdict:

 

The 1.4 version, with an improved use of the CPU and its cores, guarantees an equitable redistribution of the necessary calculations to be redirected after the GPU, also a few percentage points of the work that the GPU has to do and then enter the signal to our monitor
the decline in average use that I recorded for the GPU is around 10-12% , this gain allows a lower load for the processors and will free its performance, especially for the CPU

This translates into a smoother game, with no shuttering and lag , a larger and more stable number of FPS in their peak and low peaks and their total average.

For what I noticed the 1.4 version even in the long sessions guarantees me to play beautifully with an average closer to 60 FPS, in some maps is around 63-65 fixed FPS.
What does not happen in the current version 1.3 where my fps start to fall after the first games bringing me an average of 45-55 FPS with drops that touched the 30, making the game uncomfortable with lag and shuttering or poor synchronization between the bottom and top of my 65Hz monitor

I also tried the maps for the Grand Battles on the 1.4 test and I repeatedly got an average of 50-55 fps compared to 40-45 of the 1.3 version, the gains for me are completely significant

 

*clearly I have run tests on identical maps to have an ideal comparison, these screenshots are entirely indicative

 

1.3 

 

 

 

1.4

 

 

 

A totally necessary patch with regard to the performance of the encore graphics engine, which not all software houses do, disregarding the optimization of the games and the wallets of the players
This avoids me upgrading, at least for the moment, RAM - CPU and MOBO
A nice savings for my wallet

Really satisfied with this new development, probably the best thing contained in the 1.4 patch , I sincerely hope that those who own even worse computers than mine can have significant increases in performance

 

 



Cobra6 #2 Posted 16 January 2019 - 03:10 PM

    Field Marshal

  • Beta Tester
  • 16517 battles
  • 17,458
  • [RGT] RGT
  • Member since:
    09-17-2010

The screenshot comparison I'd chosen the same actual location on the map. Problem is that a completely flat map like Prokh renders much faster due to there being much less models (and thus textures) to draw.

 

That being said, a good performance increase is always welcome!

 

Cobra 6


Edited by Cobra6, 16 January 2019 - 03:11 PM.


ValkyrionX #3 Posted 16 January 2019 - 03:15 PM

    Captain

  • Player
  • 52677 battles
  • 2,462
  • [RDDT] RDDT
  • Member since:
    02-07-2015

View PostCobra6, on 16 January 2019 - 03:10 PM, said:

The screenshot comparison I'd chosen the same actual location on the map. Problem is that a completely flat map like Prokh renders much faster due to there being much less models (and thus textures) to draw.

 

That being said, a good performance increase is always welcome!

 

Cobra 6

 

 

as I wrote above the screenshots are entirely indicative, I have compared identical maps in the training battles in 1.3 and 1.4 and the results are those shown in the game screenshots that I chose randomly to avoid having to load +50 screenshots

in any case, best performances are welcome

 



qpranger #4 Posted 16 January 2019 - 03:17 PM

    Major General

  • Player
  • 38887 battles
  • 5,933
  • [HAMMY] HAMMY
  • Member since:
    12-25-2013
So for the Readers Digest version: will my garage screen stop lagging? :)

ValkyrionX #5 Posted 16 January 2019 - 03:19 PM

    Captain

  • Player
  • 52677 battles
  • 2,462
  • [RDDT] RDDT
  • Member since:
    02-07-2015

View Postqpranger, on 16 January 2019 - 03:17 PM, said:

So for the Readers Digest version: will my garage screen stop lagging? :)

 

potentially yes, write your hardware configuration here

qpranger #6 Posted 16 January 2019 - 03:22 PM

    Major General

  • Player
  • 38887 battles
  • 5,933
  • [HAMMY] HAMMY
  • Member since:
    12-25-2013

View PostValkyrionX, on 16 January 2019 - 04:19 PM, said:

 

potentially yes, write your hardware configuration here

 

I don't remember exactly but I determined that the hiccups are due to processor multithreading not occurring properly. Can you tell me if WG are adding a special setting for that in-game, or will it just improve by itself? Because I set everything properly in various video card and processor utilities ages ago.

ValkyrionX #7 Posted 16 January 2019 - 03:30 PM

    Captain

  • Player
  • 52677 battles
  • 2,462
  • [RDDT] RDDT
  • Member since:
    02-07-2015

View Postqpranger, on 16 January 2019 - 03:22 PM, said:

 

I don't remember exactly but I determined that the hiccups are due to processor multithreading not occurring properly. Can you tell me if WG are adding a special setting for that in-game, or will it just improve by itself? Because I set everything properly in various video card and processor utilities ages ago.

 

there will be no option in game or in your graphic driver software that allows you to enable or disable the multi-threading of the graphic engine, simply this function is present in the encore graphics engine and will order the program to render the work of the graphics engine to all the cores available in the fairest possible way, the more cores numbers are present in a processor, the better the computational work will be, the lower this number of logical processors, the lower the gain will be in terms of total performance.


This does NOT mean that even players with processors with only 2 cores do not have an increase in overall performance as far as the stable and higher stability of the FPS is concerned.

you should download the CT client in first person to understand what is going to change for your hardware

 

Edit:

I forgot that this clearly translates into better performance even for those who play WOT with laptops not necessarily for gaming


Edited by ValkyrionX, 16 January 2019 - 03:33 PM.


qpranger #8 Posted 16 January 2019 - 03:34 PM

    Major General

  • Player
  • 38887 battles
  • 5,933
  • [HAMMY] HAMMY
  • Member since:
    12-25-2013
 

Thanks, I have four cores, so should definitely see some improvement.



eldrak #9 Posted 16 January 2019 - 03:37 PM

    Second Lieutenant

  • Player
  • 51226 battles
  • 1,354
  • [GR-W] GR-W
  • Member since:
    01-27-2011

Nice comparison.

Except for putting 1.3 and 1.4 spoilers in the wrong order under CPU+GPU evaluation.



ValkyrionX #10 Posted 16 January 2019 - 03:41 PM

    Captain

  • Player
  • 52677 battles
  • 2,462
  • [RDDT] RDDT
  • Member since:
    02-07-2015

View Posteldrak, on 16 January 2019 - 03:37 PM, said:

Nice comparison.

Except for putting 1.3 and 1.4 spoilers in the wrong order under CPU+GPU evaluation.

 

yeah I know it xD

 

but the thread remain interesting imo  

14:47 Added after 5 minutes

View Postqpranger, on 16 January 2019 - 03:34 PM, said:

Thanks, I have four cores, so should definitely see some improvement.

 

with 4 cores you will absolutely have better performances 

 

certainly this also depends on the architecture of a processor and from the model of gpu and motherboard that you use, their frequency, the ram is important but not as much as many think

 

surely an amd fx 4 cores processor would do better than mine amd 965 BE being a previous model
obviously we can not even make a comparison with amd ryzen or threadripper

 

same thing can be said for the intel house processors and their different generations from the old intel core 2 duo up to the processors with 9 cores and 16 threads 


Edited by ValkyrionX, 16 January 2019 - 03:48 PM.


Srbija_Forever_38 #11 Posted 16 January 2019 - 03:49 PM

    Private

  • Player
  • 13239 battles
  • 3
  • [GARDA] GARDA
  • Member since:
    12-29-2015

hmm i didn't get any performance improvements

(fx 6300)


Edited by Srbija_Forever_38, 16 January 2019 - 03:49 PM.


ValkyrionX #12 Posted 16 January 2019 - 03:54 PM

    Captain

  • Player
  • 52677 battles
  • 2,462
  • [RDDT] RDDT
  • Member since:
    02-07-2015

View PostSrbija_Forever_38, on 16 January 2019 - 03:49 PM, said:

hmm i didn't get any performance improvements

(fx 6300)

 

the variables do not depend only on the processor imo, your model should have 6 cores if I'm not mistaken


you should check the frequencies [if I'm not mistaken your stock cpu goes to 4ghz] and it is not the top of the ranges of amd FX processors

you should check the settings of your graphic driver and try to make a comparison with identical settings as I did and try to find the performance differences between a client and another

the overclocking of some MOBO [NB and HT Link cpu multiplier or FSB] can definitely have a positive effect on gaming, even on MOBOs not born for this purpose [like mine]
if you do not know what you're touching, I strongly advise against touching settings on your bios/uefi or your graphic software

 

edit:

you probably have no real improvements because of something badly set or because of a GPU that fails to make good despite the best use of the CPU

 

what model of GPU do you have?

do you use wot 1.3 and wot 1.4 ct with the same settings?

how many fps do you have in different clients? 

are your drivers all updated?

what version of windows do you use?

 


Edited by ValkyrionX, 16 January 2019 - 04:00 PM.


Dr_Oolen #13 Posted 16 January 2019 - 04:09 PM

    Lieutenant

  • Player
  • 23509 battles
  • 1,915
  • [IDEAL] IDEAL
  • Member since:
    03-13-2012

gj,

 

Personally didnt notice any difference myself, but that is because on CT i use the improved renderer and on live i dont :D

 

So on live i get 200+ fps and on ct i have some 75-90 in 1.4. But for what its worth, it seemed more stable and smooth than before as i turned the improved renderer off because the fps was jumping around a lot and the game just didnt feel good and as responsive.



Srbija_Forever_38 #14 Posted 16 January 2019 - 04:10 PM

    Private

  • Player
  • 13239 battles
  • 3
  • [GARDA] GARDA
  • Member since:
    12-29-2015

 

View PostValkyrionX, on 16 January 2019 - 03:54 PM, said:

 

the variables do not depend only on the processor imo, your model should have 6 cores if I'm not mistaken


you should check the frequencies [if I'm not mistaken your stock cpu goes to 4ghz] and it is not the top of the ranges of amd FX processors

you should check the settings of your graphic driver and try to make a comparison with identical settings as I did and try to find the performance differences between a client and another

the overclocking of some MOBO [NB and HT Link cpu multiplier or FSB] can definitely have a positive effect on gaming, even on MOBOs not born for this purpose [like mine]
if you do not know what you're touching, I strongly advise against touching settings on your bios/uefi or your graphic software

 

edit:

you probably have no real improvements because of something badly set or because of a GPU that fails to make good despite the best use of the CPU

 

what model of GPU do you have?

do you use wot 1.3 and wot 1.4 ct with the same settings?

how many fps do you have in different clients? 

are your drivers all updated?

what version of windows do you use?

 

GPU RX 560 4GB

fx 6300 oc to 4.5 (6 cores)

Windows10 Pro 64-bit

yes i always update drivers and i use same settings

 

 

 

 



ValkyrionX #15 Posted 16 January 2019 - 04:20 PM

    Captain

  • Player
  • 52677 battles
  • 2,462
  • [RDDT] RDDT
  • Member since:
    02-07-2015

View PostSrbija_Forever_38, on 16 January 2019 - 04:10 PM, said:

 

GPU RX 560 4GB

fx 6300 oc to 4.5 (6 cores)

Windows10 Pro 64-bit

yes i always update drivers and i use same settings

 

 

 

 

 

if you have the rx560 4gb [a good card] you should go to the panel related to the game settings in the radeon software, manually look for the executable file of wot 1.4 CT which of course is called as the normal one of the 1.3 "worldoftanks.exe" and create a profile graph identical to what you use for wot 1.3, the same thing you have to do on wattman profiles regarding overclocking that are activated as soon as you launch the game on your computer, must be identical for those you use on wot 1.3

 

I have the radeon driver for the whql version 19.1.1 

 

currently on the amd website two different drivers are available for all RX500 family cards

 

version 19.1.1 and version 18.12.2

 

https://www.amd.com/...s/radeon-rx-560

 

 

 

 

 

15:27 Added after 7 minutes

View PostDr_Oolen, on 16 January 2019 - 04:09 PM, said:

gj,

 

Personally didnt notice any difference myself, but that is because on CT i use the improved renderer and on live i dont :D

 

So on live i get 200+ fps and on ct i have some 75-90 in 1.4. But for what its worth, it seemed more stable and smooth than before as i turned the improved renderer off because the fps was jumping around a lot and the game just didnt feel good and as responsive.

 

 

this depends on a discrete variable of settings that start from the bios up to the operating system, from the hardware to the settings related to the graphics drivers and many other hardware variables

certainly even if many will not notice true differences, surely use all cores in a better way than always having the core 0 to 100% even on cpu with 8+ cores is certainly something positive.

above all because it will free up the available resources of the graphic processor, whatever it is.

 

 

 

I should have written it in the initial thread but I add it here ..

clearly, if a player has [like me] a well-ventilated computer where the temperatures are always stable and below the recommended thresholds you can also think of decreasing or disabling the CPU termal throttling which lowers the frequencies of the processor to cool it down. as soon as possible.

 

this function is born to protect the hardware, but it literally kills the performances during the game.

I have completely disabled and the gain I notice strongly in the game and in the common use of my pc.

The CPU always works at the maximum frequency set [3.7ghz] without ever going down, clearly if you do not have a well dissipated system this should not be done, otherwise the operating system would crash and the BIOS automatically shut down the pc to protect it from overheating.

 


Edited by ValkyrionX, 16 January 2019 - 04:34 PM.


jack_timber #16 Posted 16 January 2019 - 06:04 PM

    Lieutenant Сolonel

  • Player
  • 40301 battles
  • 3,249
  • Member since:
    07-26-2014

Very interesting. 

Did try CT briefly last night but was laggy, ping was 90 to 120 normally on live it's 35 to 40. As a consequence I didn't notice any difference and had 2 CTDs. 

Will await an update and try again.



DeadLecter #17 Posted 16 January 2019 - 06:20 PM

    Lieutenant

  • Player
  • 32223 battles
  • 1,752
  • Member since:
    05-28-2016
These changes are very welcome for guys like me with ancient PCs. Got an i3 4170 paired with 2 gigs of 750 TI and 8 gigs of ram which is quite ancient. And I usually get above 50 FPS but it can drop to as low as 30 which gets uncomfortable. BTW I play on High settings at 1440*900. My monitor is pretty old. Kinda in for an upgrade but currently on a tight budget so that upgrade has gotta wait.

NekoPuffer_PPP #18 Posted 16 January 2019 - 07:06 PM

    Colonel

  • Player
  • 34197 battles
  • 3,991
  • [VRTC] VRTC
  • Member since:
    09-13-2013

Now all the people who found the implementation of Multi-Threading in 2019 hilarious can put a big fat sock in it.

 

This is not said all too often but...GG WG! Job well done! :great::honoring:



ValkyrionX #19 Posted 17 January 2019 - 11:05 AM

    Captain

  • Player
  • 52677 battles
  • 2,462
  • [RDDT] RDDT
  • Member since:
    02-07-2015

View PostPervyPastryPuffer, on 16 January 2019 - 07:06 PM, said:

Now all the people who found the implementation of Multi-Threading in 2019 hilarious can put a big fat sock in it.

 

This is not said all too often but...GG WG! Job well done! :great::honoring:

 

to note that at least as regards the technical side this game is really very optimized compared to other online titles even of different genres
something that not all game companies make
on this point of view the WG behaves very well indeed

kubawt112 #20 Posted 17 January 2019 - 11:50 AM

    Warrant Officer

  • Player
  • 3378 battles
  • 758
  • [-UM] -UM
  • Member since:
    07-10-2012

View PostValkyrionX, on 17 January 2019 - 11:05 AM, said:

(...)

 

They of course have their reasons, but it's indeed a very flexible game. The laptop I played WoT on in 2012 still plays the game perfectly OK (it's not a great computer and never ran the game great, fwiw).




1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users