CPU-z can read the speeds, you don't need to submit yours, but we are currently trying to figure out if "First Steps" on high game speeds are limited by CPU or RAM. Currently my results when comparing to other benchmarks tells me that CPU is really the bottleneck, but both DaveMcW and Rsesding (VERY experienced factorio users) says that RAM would be more of a limitation to the performance than CPU would in my testing, so im waiting for my shiny new ram sticks in the mail to confirm wether or not better RAM impacts performance on "First Steps"MrDoomah wrote: RAM: 8GB DDR3 (dunno where I can read the actual speed)
I didn't want people to download some map i made, so i went for what i thought to be the easiest option.sillyfly wrote:It would be interesting to see how these results differ on a map with actual things happening (maybe make a simple saved-game with a belt and inserter loop, and a few logistic bots flying around...)
This is what i get with more tests incorporated. Seems more skewed towards CPU performance for me, and that the 1333MHz block just happens to be put together with the generally weaker cpus.keyboardhack wrote:Noticed a trend with RAM speed
Obviously this could just be a coincidence.
CPU clock speeds matters only within the same CPU generations. Newer CPU generations usually gets more performance per clock.pedter wrote: Try plotting CPU speed vs. UPS instead?
That was my original idea, yes. But i decided against it as i didn't think everyone would want to download some map. I may have time to make a new test out of another map later tonight, and in that case i can also create a comparison between observed performance on First Hope vs. the observed performance in a big factory map. If you have a suggestion on what map i could use though, feel free to contributekovarex wrote:This is a good idea, but I strongly believe, that it should be tested on some specific save with huge factory instead.
a set of maps may actually be betterkovarex wrote:This is a good idea, but I strongly believe, that it should be tested on some specific save with huge factory instead.
One map that has part of everything is enough I believe.ratchetfreak wrote:a set of maps may actually be betterkovarex wrote:This is a good idea, but I strongly believe, that it should be tested on some specific save with huge factory instead.
where 1 is an empty map as a null reading
one has a ton of active belts
one has a ton of active robots
one has a ton of active biters
etc.
But then a screenshot of the performance profiling should also be included (and hoping that it's accurate enough)kovarex wrote:
One map that has part of everything is enough I believe.
CPU-z reports the internal data rates for your RAM, DDR stands for double data rate, so your bus speed is double the internal data rate. So if CPU-z reads 800MHz, then you have 1600MHz bus speed.daniel34 wrote:
EDIT: edited out the RAM speed values (MHz), I think that cpu-z reports them wrong (or I'm missing something).
Your desktop PC has much better RAM, which is extremely important in video editing. While 16GB is a lot in gaming, in video editing it is considered a bare minimum, as 8GB will push a lot of the files in you workflow over to some sideload cache on you ssd. Also it has that 750TI, and every good video editing application out there will use that to accelerate the rendering process. Hence, your desktop is still much faster for those tasks, despite its slower CPUdaniel34 wrote:I'm actually surprised that my laptop (PC 3) ranks better than my desktop PC (PC 1) because in most applications (video editing/rendering, chess) it is better than my laptop.
Is it actually playable before setting game speed to 1000?prg wrote: Factorio Version: 0.12.19
Specs:
OS: Raspbian (host), Gentoo (x86 chroot)
CPU: static qemu-i386 running on an ARM1176JZF-S, 700MHz
GPU: Mesa 11.0.6 softpipe
RAM: 448MiB (64MiB of the total 512MiB are stolen by the SoC's GPU that't not even used here)
Storage: 16GiB class 6 SD card, chroot mounted over NFS
Average UPS: 0.0
Average FPS: 0.0
Didn't yet manage to zoom in, but I doubt it would increase performance much.
Factorio on a software renderer running in an emulator on a 700MHz ARM CPU? No, no it's not playable.sillyfly wrote:Is it actually playable before setting game speed to 1000?
Code: Select all
15119.321 Factorio initialised