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

[spookywatcher]

*** Edit ***
Imperial Good and others are correct that there is too much randomness with new game starts. FPS can vary 10 FPS or more which is outside the margin of error.

So the best method is for everyone to load from a common save game and post FPS relatively soon after load (10-20 seconds ?). This prevents the RNG from straying too far after loading the save game.
Please use Pref's save games. Please do all three a new game young guns in empty space save game, the dense empire dense scenery save game, and the dense empire empty scenery and report fps for all three. You can find the links below (Hope Pref doesn't mind...YOu can also see his links to the saves on page 1 of this thread further down below). Make sure you are NOT GPU limited by turning down the resolution, the detail levels and turning OFF SSAO as this seems to be a possible cpu hit. If using a version other than 3.10 SV please state that in your post.

Hopefully this will provide info for those who want to know how a different CPU performs in X4 when looking at upgrading. I will try to update the OP with benches from others.
Also, I believe my 6700k @ 4.8ghz is doing so well because it is air gapped on a fresh win 7 install w/ only the updates needed to run X4 and nvidia drivers 445. So there are no spectre or meltdown mitigation patches nerfing it. But that is just what I believe.

Much Thanks to Pref for providing Great save games.

https://www.dropbox.com/s/h4x2lisiaajmi ... ml.gz?dl=0

https://www.dropbox.com/s/rb03x7cufwov8 ... ml.gz?dl=0

https://www.dropbox.com/s/i7uwuf6il6r50 ... ml.gz?dl=0

CPU: 5800X3D Game: 5.10 with all DLC
GPU: 3060ti
Ram: 32 3600 cl 16
Resolution: 3440x1440
Young Gun Empty: 170
Dense Empire: 26
Dense Empire Empty: 96


CPU: 12900k Version 4.2 - Much Thanks to Berni for providing the numbers. Here is his thread: viewtopic.php?f=146&t=444302
Mem: 6000 DDR5 cl36
GPU: 3090
Young Gun: 159-161 FPS
Dense Emp: 39-41 FPS
Dense Emp/Empty: 138 FPS

CPU: 5950X (no overclock) Version: 3.3 (latest non-beta as per 2020-12-19)
Mem: 3600 DDR4 (XMS enabled)
GPU: 2070Super - All settings turned down, but still at 3440x1440
Young Gun : 169 - 171 fps
Dense Emp : 26 - 31 fps
Dense Emp / Empty : 167 - 169 fps

CPU : 5900x @ stock tested on version 3.30 SV
RAM : 3600 mhz CL 16 / 32 GB Dual Rank Dual Channel with memory timings tightened
Young Gun : 152-153
Dense Emp : 27-30
Dense Emp w/ Empty : 152-155

CPU : 5900x version 4.0 beta 4 ( I believe from post it's still stock..no benefit from trying overclocks)
RAM : 3600 mhz cl 16 / 32 gb dual rank dual channel with tight mem timings
Young Gun : 154 - 155 fps
Dense Emp : 33 - 36 fps (substiantial increase running 4.0 beta 4)
Dense Emp w/ Empty : 153 -156 fps

CPU : 5800x stock version 3.3
RAM : 3600 mhz cl 18 18 18 22
Young Gun : 151 fps
Dense Emp : 23 fps
Dense w/ Empty : 158

CPU : 10700K stock version 4.0 beta 4
RAM : 3600 mhz cl 15 32 gb
Young Gun : 147 fps
Dense Emp : 36 fps
Dense Emp /w Empty : 149 fps

CPU : 10900k stock version 4.0 beta 4
RAM : 3200 mhz 2x 16gb
Young Gun : 136 - 139
Dense Emp : 38 - 51
Dense Emp w/ Empty : Appears to crash beta 4

CPU : 6700k @ 5.0 ghz (all my 6700k tests are on fresh win7 no updates and airgapped so no spectre meltdown mitigation)
MEM : 3600 mhz cl 15 heavily tuned with very tight timings
Cache Core : 4.7 ghz
Young Gun : 156 fps
Dense Emp : 29 - 30 fps
Dense Emp w/ Empty : 155 fps

CPU : 6700k @ stock 4.0 ghz boost 4.2 ghz
RAM : 2800 mhz CL 15
Young Gun : 136 - 137 fps
Dense Emp : 23 - 24 fps

CPU : 6700k @ 4.6 ghz
RAM : 2133 mhz CL 16 ( this is a 5 fps hit compared to 2800 mhz cl 15 - see next cpu bench below)
Young Gun : 137 fps
Dense Emp : 22 fps

CPU : 6700k @ 4.6 ghz
RAM : 2800 mhz CL 15
Young Gun : 142 fps
Dense Emp : 27 fps
Dense Emp w/ empty : 136 - 142 fps

CPU : 6700k @ 4.6 ghz
RAM : 3000 mhz cl 14
CPU cache core : 4.6 ghz big overclock to match cpu clock
Young Gun : 143 fps
Dense Emp : 25 - 26 fps
Dense Emp w/ empty : 138 - 144 fps

CPU : 6700k @ 4.8 ghz
RAM : 2800 mhz CL 15
Young Gun : 145 fps
Dense Emp : 26 - 27 fps

CPU : 6700k @ 4.8 ghz
MEM : 3000 mhz cl 14
Cache Core : 4.6 ghz
Young Gun : 148 - 150 fps
Dense Emp : 26 - 27 fps
Dense Emp w/ Empty : 146 - 147

CPU : 6700k @ 4.8 ghz
MEM : 3600 mhz cl 15
Cache Core : 4.7 ghz
Young Gun : 154 fps
Dense Emp : 28-29 fps
Dense Emp w/ Empty : 153

CPU : 6700k @ 4.9 ghz
MEM : 3000 mhz cl 14
Cache Core : 4.6 ghz
Young Gun : 149 - 150 fps
Dense Emp : 27 - 28 fps
Dense Emp w/ Empty : 148 - 149 fps

CPU : 6700k @ 5.0 ghz (this is showing memory bottle neck - I might get new sticks soon to try)
MEM : 3000 mhz cl 14
Cache Core : 4.6 ghz
Young Gun : 149
Dense Emp : 27 - 28 fps
Dense Emp w/ Empty : 148 - 149 fps

CPU : 6700k @ 5.0 ghz
MEM : 3600 mhz cl 15
Cache Core : 4.7 ghz
Young Gun : 156 fps
Dense Emp : 29 - 30 fps
Dense Emp w/ Empty : 155 fps

CPU : 8700 @ 4 ghz (?)
RAM : 2400 mhz cl 16 w/ loose timings
Young Gun : 121 - 125 fps
Dense Emp : 19 - 20 fps
Dense Emp w/ empty : 129 - 130 fps

CPU : 8700 @ 4 ghz
RAM :2400 mhz cl 16 w/ performance timings
Young Gun : 139 - 141 fps ( massive increase from him just setting mem higher)
Dense Emp : 17 - 20 fps
Dense Emp w/ empty : 139 - 141 fps

CPU : 7700HQ @ stock 2.8 ghz boost 3.8 ghz
RAM : 2133 mhz single channel
Young Gun : 117 fps
Dense Emp : 17 - 18 fps

CPU: Ryzen 3600 (no overclock) Version: 4.0 Beta 4
Mem: 3200 DDR4
GPU: 1070 - Ultra settings except SSAO off and VSync off, resolution 2560x1440
young guns : 132 - 134
emire dense : 20 - 22
empire empty : 139 - 143

CPU : Ryzen 3600x (PBO on) Version: 4.0 b4
Mem : 320016gb ddr4
GPU : 1080ti, 2560x1440 SSAO low, SSR off, GLOW low, rest on high and LOD 100%, no AA
young guns : 124-125
empire dense : 21-23 ( holy **** that station )
empire empty : 126-128

CPU : Ryzen 3600 @ 4 ghz
RAM : 3200 mhz cl 16
Young Gun : 132 fps
Dense Emp : 18 - 19 fps

CPU : Ryzen 3600 (temp 57 tdie 57 tjunc)
RAM : 3200 cl 16
Young Gun : untested
Dense Emp : 17 - 22 fps

CPU : Ryzen 3600
RAM : 3600 mhz cl 16
Young Gun : untested
Dense Emp : 22 - 28 (?) unclear from his post

CPU : Ryzen 3600x @ 4.4ghz version 3.3
RAM : 3200 cl 16
Young Gun : 127
Dense Emp : 20 - 21
Dense Emp w/ empty : 130

CPU : 6800k @ 4.2 ghz
RAM : 3200 mhz cl 13 (? awesome if I read that right)
Young Gun : 134 fps
Dense Emp : 22 fps

CPU : 9900k @ 5.0 ghz - This is a laptop! This is why it looks out of sorts.
RAM : 2600 mhz cl ? ( it looks like mem speed is the bottleneck here)
Young Gun : 134 fps
Dense Emp : 18 fps
Dense w/ empty : 132

CPU : 9900k @ 5.0 ghz (3840x1600 resolution) - This is a laptop 9900k
RAM : 2600 mhz cl ?
Young Gun : 104 fps (definitely now a GPU constraint with the higher res)
Dense Emp : 20 fps
Dense w/ empty : 109

CPU : 2700k @ 4.5 ghz (lower res)
RAM : 1600 mhz cl ? (DDR 3 )
Young Gun : 101 fps
Dense Emp : 11 fps
Dense w/ empty : 99 fps

CPU : 2700k @ 4.5 ghz (3840 x 1600 resolution
RAM : 1600 mhz cl ?
Young Gun : 60 fps
Dense Emp : 12 fps
Dense w/ empty : 71 fps

CPU : 10600k @ Stock
MEM : 3600 mhz cl 15
Cache Core : Stock
Young Gun : 147 fps
Dense Emp : 24 fps
Dense Emp w/ Empty : 146 - 147 fps

CPU : 10600k @ 4.9 ghz ( I was doing the same score with my 6700k @ 4.8ghz - Don't upgrade..not worth it)
MEM : 3600 mhz cl 15
Cache Core : 4.8 ghz
Young Gun : 153 fps
Dense Emp : 28 fps
Dense Emp w/ Empty : 150 - 151

CPU : 10600k @ 4.9 ghz
MEM : 3700 mhz cl 15 ( just a slight boost in my mem oc )
Cache Core : 4.8 ghz
Young Gun : 153 fps
Dense Emp : 29 fps
Dense Emp w/ Empty : 151 fps

CPU : 10600k @ 5.0 ghz
MEM : 3600 mhz cl 15
Cache Core : 4.8 ghz
Young Gun : 153 fps
Dense Emp : 29 fps
Dense Emp w/ Empty : 150 fps

CPU : 7700k @ 4.49ghz version 3.3
RAM : 2666 mhz cl 16
Young Gun : 127-130 fps
Dense Empire : 22-23 fps
Dense Emp/ empty : 120-127 fps

CPU : 7700k @ 4.49 version 4.0 beta 4
RAM : 2666 mhz cl 16
Young Gun : 126-130 fps
Dense Empire : 24-25 fps
Dense Emp/ empty : 126-130 fps



*** Don't do the below instructions Use the instructions above ***
_________________________________________________________________

Here's how to do it.
Make sure you are not GPU limited by selecting lower resolution and detail levels.
Turn off SSAO this might be a big cpu hit.
Make sure you are not FPS limited by vsync or by GPU performance
Start a new game...Young Gun.
Undock and fly south until you are well away from stations (further than scan range) and other ships. Come to a stop. As long as you are not GPU limited this should give "Universe Speed" limited by CPU perf
Note FPS then and then again at around 2.5 mins (gives time for universe to settle and or ramp up a bit)

I'm really interested in CPU performance (as this game is severely cpu bottle necked). Thinking about upgrades (intel 10 series and ryzen 4000)
For me this is very reproducible so this is best way I could come up with.
Thx for posting.

My results:

6700k @ 4.6 ghz - hyperthreading on
16 gig DDR4 @ 2800mhz xmp
Win 7 - v3.1 SV

130 FPS once I got south
126 FPS after about 2.5 mins

[Imperial Good]

Start a new game...Young Gun.
Undock and fly south until you are well away from stations (further than scan range) and other ships. Come to a stop. As long as you are not GPU limited this should give "Universe Speed" limited by CPU perf

This is not consistent. Universe generation involves RNG and hence the placement of objects may vary and hence the interactions with those objects may vary.

Would be better to provide a save game which one just loads and measures from. Should include Split Vendetta DLC as that expands the universe a lot so is more CPU demanding.

One must also state the version used. There may be subtle performance differences between current live and beta 3.20.

[spookywatcher]

Start a new game...Young Gun.
Undock and fly south until you are well away from stations (further than scan range) and other ships. Come to a stop. As long as you are not GPU limited this should give "Universe Speed" limited by CPU perf

This is not consistent. Universe generation involves RNG and hence the placement of objects may vary and hence the interactions with those objects may vary.

Would be better to provide a save game which one just loads and measures from. Should include Split Vendetta DLC as that expands the universe a lot so is more CPU demanding.

One must also state the version used. There may be subtle performance differences between current live and beta 3.20.

** sigh **

It was reproducible for me even across platforms. eg. My linux install was almost exactly the same results. And if I uninstall and reinstall in either win 7 or in linux and re-do my instructions I get again...almost the exact results. Within 3 fps. It is reproducible time and time again. I've tested it quite a bit.

Also...I did post which version. See OP. It was Win 7 v 3.1 SV

Not looking for debating semantics on whether it is consistent. It is consistent enough for me and my curiosity. I'd love to see someone with a 7700k post as that is my quickest upgrade (no new mobo). I'd love to see all Ryzens post. I've love to see 9900k stock and overclock.

So if you have a score I'd like to see it. Whether someone thinks it's not a good benchmark is irrelevant. I do agree that a savegame will be more consistant. I'll try to get it.

[pref]

Not sure this would tell you much. How often a blank screen can get rendered barely has to do anything with cpu.

You should check this with a save that has a big empire going in the background because without stressing the cpu there won't be a "bottleneck".
Ime the biggest factor is what the game has to render in 3D, turret aim and pure object count. I have quite a few assets and fps is nice when im in normal space.

Otherwise i think i had around 160~180 fps without msaa and vsync in quiet places with an i8700 and a gtx1070.

[spookywatcher]

Not sure this would tell you much. How often a blank screen can get rendered barely has to do anything with cpu.

You're wrong. In x4 if you are not GPU bound then you are CPU bound. And in the "beginning" as the "universe" is being generated is the most consistent test I can come up with.

When I clocked my 6700k @ 3.6 ghz and did the same test my FPS was 123 and 116 after 2.5 mins.
When I clocked my 6700k @ 4.0 ghz and did the same my FPS was 127 and 122 after 2.5 mins (after many tests...it was always within 3fps) and with cpu clocks this scales linearly. So yes...when you are cpu bottlenecked going to blank space will show how the cpu is the bottleneck.
What I can't simulate is different IPC's for the different CPU's. Or differences between two CCX or single CCX Ryzens.

So you are wrong. But much appreciate the FPS you gave with your 8700. Was it a new game start young gun and immediately fly to open space without gpu bottleneck?

[euclid]

I tend to agree with pref, meaning that it won't tell you much about cpu load. Anyway, stationary in empty space I get about 120 fps (XFX AMD Radeon R9 390 8192 MB GDDR5,i5-2400 @3.10GHz).

Cheers Euclid

[pref]

In x4 if you are not GPU bound then you are CPU bound.

o,O

[Imperial Good]

It was reproducible for me even across platforms. eg. My linux install was almost exactly the same results. And if I uninstall and reinstall in either win 7 or in linux and re-do my instructions I get again...almost the exact results. Within 3 fps. It is reproducible time and time again. I've tested it quite a bit.

Then someone does it and happens to get some strange RNG seed where large fleets start out in combat or other edge cases occur and is 10 frames per second lower which is outside margin of error. This is why a save file should be used since it minimizes variables. Otherwise one would need to repeat the test multiple times and then use statistics to obtain consistent results to reduce the impact of the variables or prove they are independent.

It was Win 7 v 3.1 SV

My apologies, I thought that was the version of the operating system like Windows 10 - v1909 18363.836. Usually one lists the software version separately to avoid such confusion.

You're wrong. In x4 if you are not GPU bound then you are CPU bound.

Or memory bandwidth/performance bound, which is what quite a few people report may be the case. Someone posted that they gained more performance from using faster memory than an extra 0.1 MHz on their core i9 9900k. This may well be the case for you given how an 11% increase in clock speed only increased frame rate by ~3%. Assuming you were not GPU bottlenecked in that scene of course.

Or differences between two CCX or single CCX Ryzens.

Someone reported that forcing X4 onto a single core cluster chiplet does yield better performance than allowing it to run between both. I have yet to verify this.

Within core cluster chiplets it likely is detrimental to limit X4 to only 1 complex since that would restrict it to just 3 or 4 CPU cores. It probably has scaling benefits beyond that, especially with the low core count complexes. Extensive testing would be required as it is possible that different activities of X4 scale better with core count than others.

[Old Drullo321]

As stated above you have to reduce the amount of randomness as much as you can. So everyone start a new save and just following basic instruction is likely less comparable then loading one or more predefined saves whith different scenes (e.g. in a station, watching a fight, etc.) Even then there is likely randomness but that would be a good point to step further into details.

[pref]

Or memory bandwidth/performance bound

While staring at the background stars? The above scenario needs nothing from cpu, hardly any data needs to be transferred for render either.
It just tells you how quick the gpu can render a scene in x4. There needs to be some background load to have any noticeable effects on the fps, plus a control check for gpu render speed like OP suggests then one could compare the differences between these across different hardware.
Still the result is easily predictable, with decent rig the game runs nice unless the situation gets too dense where even a top pc dies. The situation is gamestyle dependent so there goes any objectivity, people who play with 100s of ships in battles will be hurt by performance issues much more then someone who operates with just 10 ships at a time.

And as the sticky says, single thread performance > thread count so that settles the amd/intel debate.

[SirNukes]

While staring at the background stars? The above scenario needs nothing from cpu, hardly any data needs to be transferred for render either.
It just tells you how quick the gpu can render a scene in x4. There needs to be some background load to have any noticeable effects on the fps,

X games aren't like Terraria or Minecraft or whatever, where action only happens around the player. Starting an X4 game, there are around 6000 ships and 1000 stations populating the galaxy, constantly running their ai script logic, movements, enemy visibility checks, trade deal searches, ware price adjustments, etc. That is a very large background load on the cpu. If you play around with starting ship counts, it is easy to see fps shoot up to 250+ (with <1k ships) or tank down to <30 fps (with 20k ships), when measuring in empty space.

[Imperial Good]

While staring at the background stars? The above scenario needs nothing from cpu, hardly any data needs to be transferred for render either.

Yet there was not a linear increase in frame rate with GPU frequency. Hence either there is a GPU bottleneck, or some memory metric (latency, bandwidth) is limiting CPU performance.

And as the sticky says, single thread performance > thread count so that settles the amd/intel debate.

Currently for an Intel CPU to perform better than a Zen2 CPU it will need to be 12nm running at a boost clock of roughly 4.7 GHz or higher that it can sustain indefinitely as well as have memory operating at the highest allowed stock frequency or better. Modern Intel processors will run X4 better than AMD Zen2 ones but the difference is not huge and anything sub optimal will quickly shrink that gap.

[pref]

Starting an X4 game, there are around 6000 ships and 1000 stations populating the galaxy, constantly running their ai script logic, movements, enemy visibility checks, trade deal searches, ware price adjustments, etc. That is a very large background load on the cpu.

Not really, background tasks and render works on different cpu threads. And these numbers are low considering how cheap OOS calcs can get. If there is a 1sec delay on whatever background task the user won't notice it anyway. Problems usually start when there is a lot to render or there is a lot of cpu calculations that need to happen realtime. Collision, aim and hit reg etc.

As long as the render thread is not hindered (which shouldn't happen even with a relatively high background task load on a modern cpu) and is not waiting for collision detection or whatever you will get proper fps. Those background tasks will just wait or run independently on another core most likely.

Modern Intel processors will run X4 better than AMD Zen2 ones but the difference is not huge and anything sub optimal will quickly shrink that gap.

Yes, marginal difference. Cooling would probably matter much more then which cpu the player has (above recommended specs).


I can check a new start and my old save in a quiet place to get an fps measure. Became curious, think fps improved recently with big asset count. I'll upload 2 saves, a big one and a fresh - both in an empty area. Then anyone can see for themselves on a fixed state.

[pref]

So i got the saves (wait 10s or so after load so fps can settle if it wants to, watched a 10-20sec period):


My results are
Young gun @ starting station: 80-100
Young gun @ emptyness: 121-125

Empire @ emptyness: 129-130
Empire @ dense area: 19-20, occasional lag spikes

And cheers to ES, that's how well they optimised their code. I had a feeling they worked on this since 1.x and the numbers speak for themselves.

I had all settings on highest except screeen space reflections and ssaa - those i turned off entirely.

rig:
i7 8700, 16G DDR4 NB freq 4ghz, GTX 1070, 1080p

I just put in the dense empire for my selfish reasons, if anyone can render that @ stable 20+ please let me know your hardware.

Edit: don't wait longer then necessary, rng can cause our paths to deviate and bring extra noise. Added screen resolution.
Use gfx settings so you get 100+ fps @ the young gun empty save, only the difference is important afterall.

[Socratatus]

This is not consistent. Universe generation involves RNG and hence the placement of objects may vary and hence the interactions with those objects may vary.

Would be better to provide a save game which one just loads and measures from. Should include Split Vendetta DLC as that expands the universe a lot so is more CPU demanding.

One must also state the version used. There may be subtle performance differences between current live and beta 3.20.

I didn`t realise there was an RNG thing going on. Love randomisation. What, for example, is randomised? Is it just lock boxes and ships or more? I know the AI Factions might be doing their own thing, but what else?

[Imperial Good]

I didn`t realise there was an RNG thing going on. Love randomisation. What, for example, is randomised? Is it just lock boxes and ships or more? I know the AI Factions might be doing their own thing, but what else?

Types and positions of lesser stations. Important stations like trading stations and shipyards/warfs usually are in the same locations but lesser ones like Energy Cell or Claytronic factories can be placed randomly around in sectors. I think the same applies to the initial fleets to some extent. This gives the universe a familiar feel while also adding some variation in that one cannot depend on lesser stations being where they were in your last playthrough. Claimable ships are not random and will always be in the exact same location.

[pref]

what else?

Station and fleet positions and layout. Guess those are the most important.
But target selection can also involve rng and player location change can have an enormous impact on events as well so you can easily go different routes from the same state without directly interacting with other ships or stations.

[spookywatcher]

Or memory bandwidth/performance bound

While staring at the background stars? The above scenario needs nothing from cpu, hardly any data needs to be transferred for render either.
It just tells you how quick the gpu can render a scene in x4. There needs to be some background load to have any noticeable effects on the fps, plus a control check for gpu render speed like OP suggests then one could compare the differences between these across different hardware.
Still the result is easily predictable, with decent rig the game runs nice unless the situation gets too dense where even a top pc dies. The situation is gamestyle dependent so there goes any objectivity, people who play with 100s of ships in battles will be hurt by performance issues much more then someone who operates with just 10 ships at a time.

And as the sticky says, single thread performance > thread count so that settles the amd/intel debate.

Again...I'm sorry but you are wrong.

The above scenario needs nothing from cpu

The above needs nothing from the GPU. And it's waiting on the CPU to calc it's universe stuff. Otherwise I'd have 300+ fps.

At the opening menu with my graphics turn way down no AA vsync off and 1440p I'm getting 280-300 fps and the GPU is maxed at 100% (It's NOT waiting on cpu to calculate background universe stuff and CPU only shows 17% util). When it cuts to the quote of a notable person I get over 5000 fps (No cpu load and no overhead rendering triagles). During "loading scene" of game I get around 2400-3000 fps (a cpu core (s) is (are) loaded a bit so not getting the 5000 fps) . And during all this time GPU is being max'd at 100 %. Why? Because it's not waiting on the cpu. With hyperthreading off (so I can see TRUE core utilization) the cpu is showing very little load. With hyperthreading on...a single thread max'd out at 100% will not show the core at 100% necessarily.

When I actually start game play and head to empty space (with nothing really to render) I'm getting 130 fps and a GPU utilization of ONLY 28%. Why? Because one core is pegged at 100%. It's not pegged at 100% because it's rendering anything. It's pegged at 100% calculating Universe stuff. And THIS is the cpu bottleneck of X4. If there was NO calculating of background universe stuff I'd be getting 300 fps like in the opening menu scene.

And this is why I want a way to measure CPU performance in this game because as Imperial Good mentioned....it's not scaling perfectly and I want to see if it's even worth going to a 9900k or 3950x or 10th gen or Ryzen 4000 etc.

It amazes me. Even in my OP I state "For me this is very reproducible so this is best way I could come up with" It doesn't matter if it's perfect or not. It would give me the best incite (for the most part) on reasonableness of upgrade path.

Normally whenever I see someone say post your benches...and please do it this way. I'm happy to do it! Even though I know that this metric is not taking into account all variables.

Just load a new Young Guns game..fly south....make sure you're not GPU bottlenecked and report the d*mn fps. It's simple.. instead of internet jousting and being wrong.

On Your new Young Guns save loaded I'm getting 142 fps. Way more than my normal 130 fps. So Imperial Good is correct. In order to get the best chance of comparison it needs to be a load from a savegame.

[pref]

On Your new Young Guns save loaded I'm getting 142 fps.

And that explains why this test alone is worthless. Check the empty empire game as well, that could provide some info how our cpus can handle background load.

[spookywatcher]

On Your new Young Guns save loaded I'm getting 142 fps.

And that explains why this test alone is worthless. Check the empty empire game as well, that could provide some info how our cpus can handle background load.

It's absolutely not a worthless test. I agree everyone checking from a common save game is the best way.

In your Dense save game I'm getting 27 fps with a GPU util of 28% and a CPU core is pegged at 100%. In your new Young Guns I was still low GPU util and CPU core (or thread) was still pegged at 100%. It's a good test. Whether there is small amount of background calc or LARGE background calc doesnt matter as long as the CPU is pegged and not the GPU. I don't care if everyone uses the New game save game and loads it or the Dense game save game.

As long as the GPU is not max'd out and a core (or thread if hyperthreading on) is max'd at 100% it's a good test. It's even better as Imperial Good suggested that everyone load a common save and make sure you're graphics are turned down.

Edit to add **

I'll be editing the OP to ask anyone who wants to show CPU perf numbers to use Your save games...the Young Gun new load and the Dense. I think both are great measures and as long as everyone uses the same AND takes the FPS soon after load (to prevent the RNG from straying after load).

So again my scores for

Young gun @ emptyness: 121-125

142 fps

for

Empire @ dense area: 19-20, occasional lag spikes

27 fps

6700k @ 4.6 overclock
16 gig 2800mhz xmp