Post your FPS & CPU's scores - instructions how to measure included - 5950x + 12900k + now 5800x3D!!

[YorkRoss]

Game ver: 5.10 with all DLC
Young Gun Empty: 135 Instant, 50 min, 131 avg, 135 max, 93 1% low, 58 0.1% low
Dense Empire: 21 instant, 2 min, 19 avg, 23 max, 1 1% low, 1 0.1% low
Dense Empire Empty: 60 instant, 50 min, 67 avg, 79 max, 34 1% low, 11 0.1% low
Game ver: 3.3 SV
Young Gun Empty: 157 Instant, 151 min, 157 avg, 160 max, 127 1% low, 102 0.1% low
Dense Empire: 21 instant, 19 min, 23 avg, 25 max, 8 1% low, 5 0.1% low
Dense Empire Empty: 157 instant, 149 min, 157 avg, 161 max, 111 1% low, 65 0.1% low
Game ver: 4.0 SV/no COH
Young Gun Empty: 156 Instant, 140 min, 151 avg, 156 max, 124 1% low, 90 0.1% low
Dense Empire: 26 instant, 19 min, 26 avg, 28 max, 7 1% low, 5 0.1% low
Dense Empire Empty: 155 instant, 129 min, 148 avg, 157 max, 95 1% low, 79 0.1% low

I was able to run the 5800X3D today, undervolted using the PBO2 tuner for -20 offset on all cores. I am also using 32GB of 3600Mhz CL 14-15-15-35 1T RAM. Here are my results:

Game ver: 5.10 with all DLC
Young Gun Empty: 172 Instant, 160 min, 171 avg, 174 max, 139 1% low, 68 0.1% low
Dense Empire: 26 instant, 10 min, 24 avg, 29 max, 2 1% low, 2 0.1% low
Dense Empire Empty: 54 instant, 43 min, 52 avg, 55 max, 26 1% low, 10 0.1% low

Game ver: 3.3 SV
Young Gun Empty: 180 Instant, 178 min, 181 avg, 185 max, 153 1% low, 82 0.1% low
Dense Empire: 30 instant, 22 min, 30 avg, 32 max, 4 1% low, 4 0.1% low
Dense Empire Empty: 177 instant, 175 min, 178 avg, 181 max, 128 1% low, 80 0.1% low

Game ver: 4.0 SV/no COH
Young Gun Empty: 182 Instant, 175 min, 180 avg, 183 max, 156 1% low, 137 0.1% low
Dense Empire: 36 instant, 28 min, 35 avg, 38 max, 20 1% low, 9 0.1% low
Dense Empire Empty: 182 instant, 164 min, 181 avg, 186 max, 124 1% low, 94 0.1% low

One thing i've also noticed is that the Main Menu FPS is also different vs my old 5800X. For the 5800X, it gets 160-165fps in all versions. However, for the 5800X3D, its actually different for each version tested. In 3.3 i get 520+, 4.0 gets 430+ and in 5.10 i get 215+. Not particularly important but just something interesting to me how it drops with later major patches/DLC added.

The only save that didnt see the improvement was the 5.10 ALL DLC version of the Dense Empire Empty save but that's probably down to the save being broken due to COH or ToA

[Imperial Good]

The only save that didnt see the improvement was the 5.10 ALL DLC version of the Dense Empire Empty save but that's probably down to the save being broken due to COH or ToA

I think someone reported that the "empty" space you are meant to load into is no longer empty in these conditions. This means that DLC or patch content is spawning nearby which could have random variance. As such this test likely has a large run to run variance which could explain your observation.

The slower main menus could be due to an increase in production value in more recent DLCs. I think the latest one for Tides of Avarice has very demanding effects like volumetric fog and the like which did not exist in older versions.

[YorkRoss]

The only save that didnt see the improvement was the 5.10 ALL DLC version of the Dense Empire Empty save but that's probably down to the save being broken due to COH or ToA

I think someone reported that the "empty" space you are meant to load into is no longer empty in these conditions. This means that DLC or patch content is spawning nearby which could have random variance. As such this test likely has a large run to run variance which could explain your observation.

The slower main menus could be due to an increase in production value in more recent DLCs. I think the latest one for Tides of Avarice has very demanding effects like volumetric fog and the like which did not exist in older versions.

Makes sense. As to the FPSes being different for 5800X3D for each version but staying the same for 5800X is probably because, due to the stronger CPU/Mem performance combined with the new effects introduced in COH and ToA, the Main menu switched to being GPU bound rather than CPU bound under the 5800X3D system while its still CPU/Memory speed and latency bound for the 5800X system. But i could be misunderstanding how those new effects work

Another thing that's broken in the Dense Empire "Empty" save is that it spawns your ship inside one of your wingman's fighters, on that last run i had no control over my ship as it was constantly colliding with said wingman's ship and causing me to roll and turn slowly even with no input.

[jkflipflop98]

Would be neat-o if we could just get a built-in benchmark tool. Nothing fancy just run the same level everytime so we can compare on truly equal footing.

[vadiolive]

I suggest this alot times , its win win but i guess is time to wait X-22(project name)
Like Starcitzen https://robertsspaceindustries.com/telemetry

[Metran]

Map performance can be improved by turning off some of the overlays. For example turning off friendly orders and player orders so it only shows order lines for selected ships can drastically improve map performance.

The main way to increase performance while viewing the map is to disable all your satellites that way you drastically reduces the number of objects the game is required to visibly keep track of, until you get to the point where you've bought all the trade subscriptions from factions.

[spookywatcher]

I finally kissed 180 fps in game on a test save of mine.

TLDR:

4 sticks of ram (at any speed) made way more of a difference than 2 sticks even clocked at 4000 CL 18 and ridiculous tight timings. Even 4 sticks at 3200 mhz cl 14 with same ridiculous tight timings was better than 2 sticks at 4000 and fclck at 2000 for 1:1.

Now for the details:

I am able to get 4 sticks at 3800 mhz cl 14 with same ridiculous tight timings but it looked like there was a little performance regression and same with 3600mhz. So I settled on 3400mhz cl 14. This is on samsung b-die at 1.45 v and soc of 1.07 v

Now...when it comes to the 5950x there is some weird behavior going on when trying to get above 4.7 ghz. My better sample of my 2 5950x can do 5.0 ghz on ccd0 and 4.9 ghz on ccd1 with aggressive overclocks. So I set that. But in game the "effective clocks" never got above 4.7 ghz. I'd see 5.0 ghz during loading etc...but in game I was hitting brick wall at 4.7 ghz effective clocks (clock stretching).

Next... X4 only uses a few cores. But it does use a few cores. So YOU WILL NEVER GET a benefit of a single core boost in PBO! Let me say that again. In game...if your PBO will do 5.1 ghz in single thread cinebench...YOU WILL NOT GET that benefit in X4 (except maybe in loading). Period...full stop...fight me on it. That's because X4 DOES use 3-4 cores. So no single core boost. AND!!! Even with my cpu clocked at 5.0 ghz I never saw an effective clock above 4.7 ghz anyways!

So...

I disable PBO, curve optimizer etc etc.

I leave everything default settings except for XMP (DHCP) and I undervolt the 5950x

So my settings:

XMP 3400 mhz cl 14 (with tightest timings possible) @ 1,45 v on samsung bdie 4 sticks of 8gb
FCLCK at auto (so running at 2000mhz)
soc volts at 1.07 (that might be a little high...I could prolly get away with 1.05)
CPU volt offset to -.0825

This gave me my highest FPS yet to date. Rock stable and nice and cool running. The 5950x still hits at 4.7 ghz effective clocks yet with WAY less voltage and much cooler temps.

Hope it helps someone...and I can't wait to get my hands on 13th gen intel and Ryzen 7000 to test

[Imperial Good]

So I set that. But in game the "effective clocks" never got above 4.7 ghz. I'd see 5.0 ghz during loading etc...but in game I was hitting brick wall at 4.7 ghz effective clocks (clock stretching).

I think in the past people came to the conclusion the reported "effective clock" being so low or seeming limited is due to pipeline stalls. Logically these pipeline stalls would likely be cache misses which likely need memory access to resolve. Pushing frequency beyond the effective clock will give very little returns.

This was observed on Intel processors in the past as well where an extra 200+ MHz odd core frequency from an aggressive i9 overclock made as good as no difference to performance while some small improvements to memory timing and frequency did.

[hxsgame]

......Hope it helps someone...and I can't wait to get my hands on 13th gen intel and Ryzen 7000 to test

13th gen intel and Ryzen 7000, They're all ready -)

[burger1]

......Hope it helps someone...and I can't wait to get my hands on 13th gen intel and Ryzen 7000 to test

13th gen intel and Ryzen 7000, They're all ready -)

I5-13600k with ddr5 should be close to the i9-12900k score?

Probably 1 or 2 more gens and ddr5 improvements before 60 fps at dense empire.

[hxsgame]

I5-13600k with ddr5 should be close to the i9-12900k score?

Probably 1 or 2 more gens and ddr5 improvements before 60 fps at dense empire.

When 13600k At the default frequency, likely to be this. and 13600k may overclock to 5.8GHz, Looking forward to seeing the test results soon

[Imperial Good]

I5-13600k with ddr5 should be close to the i9-12900k score?

Probably significantly higher due to the extra cache. At least in some of the tests.

[hxsgame]

Probably significantly higher due to the extra cache. At least in some of the tests.

Curious and Consult about L3 caching.

how much cache, well be enough temporary storage the data, in the some of typical complex scene of X4 ?

Is it possible that such a reference table exists?

[spookywatcher]

......Hope it helps someone...and I can't wait to get my hands on 13th gen intel and Ryzen 7000 to test

13th gen intel and Ryzen 7000, They're all ready -)

Yeah... sorry. I was going to test them all. But I've given up on X4 and the Devs. I bring up extremely valid and game breaking bugs (I think they should have my posts on speed dial ) but it's crickets from them. So I've moved on to a game I gave up on almost 10 years ago but is now actually quite engaging, fun and exquisitely gorgeous. Really enjoying it.

[hxsgame]

... So I've moved on to a game I gave up on almost 10 years ago but is now actually quite engaging, fun and exquisitely gorgeous. Really enjoying it.

may be Rebel Galaxy

[Wanabe]

I7 7700K
AMD 6800, Windows 10
GPU Driver 22.8.2
I7 7700k @ stock
Spectre & Meltdown Protection temporarily disabled for tests, untested performance impact if enabled

Version 5.10, Split Vendetta only

2X8GB 3200 15-17-17-35-2T
205 FPS Main menu

Young gun: 111-133, average around 129

Dense Empire:
Test 1: 6-16 average 13, 27 FPS while paused
Test 2: 12 minimum otherwise same results

Dense Empire Empty: First and Second test: 4 FPS with only decimal point deviations

Version 3.30, Split Vendetta only

2X8GB 3200 15-17-17-35-2T
205 FPS Main menu

Young gun: 131-138, average around 134, 382 FPS while paused
Dense Empire: 14-20, average around 18. 29 FPS while paused
Dense Empire Empty: 128-135, average around 130. 270 FPS while paused

4X8GB 3200 16-18-18-36-2T, Dual Rank Disabled (Interleaving OFF)

Young gun:
Test 1 (done after testing both dense empires without fully relaunching game): 126-138, average around 128, 350 FPS while paused
Test 2 (relaunched game): 133-138, average around 135, 375 FPS while paused

Dense Empire: 12-22. average 20, 30 FPS while paused. Dropped to 12 only when the ship bumped the player ship, after and before that the minimum only went down to 19.
Dense Empire Empty: 131-136, average around 132. Fluctuations of 265-272 FPS while paused

4X8GB 3200 16-18-18-36-2T, Dual Rank Enabled (Interleaving ON)

Young gun: 132-139, average around 135, 389 FPS while paused
Dense Empire: No noticeable change from the previous test with interleaving disabled
Dense Empire Empty: 137-142, average around 139. Fluctuations of 295-303 FPS while paused

4X8GB 3200 15-17-17-34-2T, Dual Rank Enabled (Interleaving ON)
214-240 FPS Main menu

Young gun: 133-137, average around 135, 397-401 FPS while paused
Dense Empire: Marginally higher than previous test but decimal differences. 31 FPS while paused
Dense Empire Empty: 137-142, average around 139. Fluctuations of 300-309 FPS while paused

These FPS values were taken using the games built-in FPS monitor over "some time". I didn't look into using software to automatically record all the FPS values over a fixed time period for proper accuracy which would certainly help with some of the inconsistencies with recording it this way. WIth that in mind I hope the above is still somewhat useful information.

All 5.10 tests showed lower FPS than 3.30 using these old savefiles. Clearly "Dense Empire Empty" was somehow broken in the 5.10 tests I tried, which is most likely a bug due to the age of the savefiles. Egosoft have shown through their various updates to X4 and their previous titles that they have a trend to noticeably improve performance from each major game update with some smaller improvements in between the bigger releases. I think it is likely that the FPS reductions is due to it being an older save that perhaps the game can't properly account for. I wonder if after the next game update from Egosoft is released if it's worth making/trying out a new set of saves without some lingering "old game" stuff going on.

I think all the Low Attention calculations are in a fairly good place nowadays, at least for the current Vanilla game setup. Improvements are always ongoing, naturally. However at this stage I'd like to see some more work on finding ways to reduce the cost of High Attention. Busier areas in general cost a lot in terms of performance as it stands unfortunately. Even pausing the game during these busy scenes can result in "low" FPS.

[magitsu]

Asteroid fields + big battles seem the most obvious reasons for chugging. I'd avoid in the future having excess of both at the same place, i.e. don't make another Second Contact Flashpoint. Matrix 451 is also bad, even with 5800x3d.

[Imperial Good]

I think it is likely that the FPS reductions is due to it being an older save that perhaps the game can't properly account for. I wonder if after the next game update from Egosoft is released if it's worth making/trying out a new set of saves without some lingering "old game" stuff going on.

There are many factors. Most obvious of which is that more DLCs will mean lower frame rate due to bigger universe to simulate. The other part has to do with migration changes which might temporarily impact performance and affect the performance profile of an area on load. Lastly there are small changes that might impact the performance profile of an area, for example volumetric fog.

[Wanabe]

There are many factors. Most obvious of which is that more DLCs will mean lower frame rate due to bigger universe to simulate. The other part has to do with migration changes which might temporarily impact performance and affect the performance profile of an area on load. Lastly there are small changes that might impact the performance profile of an area, for example volumetric fog.

Oh for sure, there are so many variables that can effect the performance from one patch to another. Even with keeping the same DLC enabled (Split Vendetta) to try to keep the DLC aspect the same there are still all the other various changes in effect. AI probably has some improved capabilities, low attention can account for more factors and/or is more accurate in places. The updates themselves add in some extra ship jobs for newer story-line elements regardless of DLC presence and also the new Scrap collecting NPC ships. All the possibilities you mentioned too of course etc.

I think Egosoft have done a fantastic job on making the game run as well as it does, especially when accounting for the overall heavier universe activity via DLCS and updates. They have not only added in more stuff going on simultaniously but found ways to improve the details of the simulation (such as better AI decisions) along the way, while also offsetting the performance impacts as much as possible with some heavy optimization passes.

Version 5.10, Split Vendetta only
4X8GB 3200 15-17-17-34-2T, Dual Rank Enabled (Interleaving ON)
214 FPS Main menu

Young Gun: 117-135 average 134, 400 fps while paused
Dense Empire: 13-19 average 17, 29 fps while paused
Dense Empire Empty:
Before moving a tiny amount: 4FPS, 94 FPS while paused
After Collision: 55-73, average 71, 104 FPS while paused

It's possible that the frame-rate issues from Dense Empire Empty could be due to collision detection. Moving a tiny amount forward caused my ship to collide with something and looking in third person there was another ship extremely close to the player ship. I think the "after" numbers won't be particularly comparable to anything though because of the random factors which occur such as where the ship ends up facing due to when the player accelerates forward and how much by. The game may have a slight random factor to the force of collision too.

[burger1]

version 5.1 all dlcs
i7-13700k
1920x1080
ddr5 5600 cl36 2x16 single rank sticks could probably easily add +10% with better ram
1660 gtx - didn't seem to be limiting fps at the settings I used but does in asteroid fields at stock settings gpu hits 98%+ usage can fix that though for 60 fps fsr etc....
win 10 pro 64bit - fresh install
disabling e cores in task manager didn't seem to change stuff much

young gun empty 150-158 fps

dense empire 28-38 fps

empire empty sector after moving out of the ship that got spawned on my ship 120-133 fps