Jump to content

FM17 Performance Benchmarking Thread


Recommended Posts

10 hours ago, yugo23 said:

After reading AMD Ryzen reviews I am even more excited to see how they will perform in FM. They could even top the table. 

I suspect they won't based on the low single thread performance 

Link to post
Share on other sites

  • Replies 217
  • Created
  • Last Reply
On 3/9/2017 at 08:42, fabioke said:

The proof of the pudding is in the eating, but I guess you're probably right.

I'd love to be proved wrong though, the pricing is really competitive and i'll be looking at a new build pretty soon.  Hopefully the next wave of AMD chips with less cores will prove easier to overclock and better suited to the game

What i'd also like to see is how the Intel Core i3-7350K fares in this type of test, it's massively overclockable and also massively affordable at just over £150

Link to post
Share on other sites

I'm using my laptop for FM17 (planning on building a Ryzen 5 when it is released though), and haven't seen any benchmarks for similar system so here is one:)

CPU: Intel Core I5-7200U

CPU Frequency: 2500 MHz

RAM: 8GB DDR3 

OS: Win 10 64-bit

Storage: SSD

Test A

Time: 9 min 23 sec

Test B

Time: 15 min 47sec

 

a bit slow i must admit, so i'll go for steam in home streaming as soon as i get hold of a ryzen chip:)

Link to post
Share on other sites

2 hours ago, lahara said:

 

I'm using my laptop for FM17 (planning on building a Ryzen 5 when it is released though), and haven't seen any benchmarks for similar system so here is one:)

 

Thanks for testing, I added your results to the table in the first post. :thup:

Link to post
Share on other sites

Figured it'd be interesting to see how an old, slightly overclocked, CPU performed. Here are my results.

CPU: i7-870 @ 3.34 GHz

Ram: 16GB DDR3 @ 1333 MHz

OS: Win 10 64-bit

Storage: HDD

Ran tests with and without hyperthreading enabled.

Test A with hyperthreading

5 min 15 sec

Test A no hyperthreading

5 min 20 sec

Test B with hyperthreading

10 min 13 sec

Test B no hyperthreading

10 min 25 sec

Link to post
Share on other sites

5-year old system still hanging in there.  Could be better though:

CPU: Intel 3930K

CPU Frequency: 4200 MHz (OC)

RAM: 32GB DDR3 1866 MHz CL10

OS: Win 10 64-bit

Storage: HDD

Test A

Time: 4 min 00 sec

Test B

Time: 7 min 36 sec

Link to post
Share on other sites

CPU: Intel i3 4160

CPU Frequency: 3600 MHz

RAM: 4GB DDR3 1333 MHz single channel

OS: Win 1064-bit

Storage: HDD @ 7200rpm

Test A

Time: 5 min 17 sec

Test B

Time: 10 min 5 sec

 

Anyone got a Pentium g4560 to test it ?

Link to post
Share on other sites

Results for ryzen:

 

CPU: AMD Ryzen R7 1700X

CPU Frequency: 3.4/3.8 GHz

RAM: 4GB DDR4 single channel

OS: Win 10 64-bit

Storage: HDD @ 5400rpm

Test A

Time: 5 min 22sec

Test B

Time: 11 min 20sec

 

Much worse than expected, probably because of ram. The ram is being used as stand in.

I may upload 12gb ram results. I will also upload 32 gb ram results in a couple of weeks and i will upload overclocked results when i upload those. unless someone else beats me to it.

Link to post
Share on other sites

@will221996 that is quite slow but I think the game really doesn't make use of all those cores which is a real shame for a game that is so statistically heavy, hopefully the next couple of versions of the game can make real use of lots of RAM and multiple cores.  Still think that Ryzen is a phenomenal bit of kit and represents outstanding value for money under the right circumstance.  looking forward to your updated results.  i'd be interested to see if the game uses all cores if you can monitor that during the test? 

Link to post
Share on other sites

yeah Ram is an issue there but the game is heavily biased towards single core performance in the main.  To be honest though i've never seen the Ram go over 6gig ever even though i've got 16gig.  They promised us a truly 64 bit game this year but clearly haven't delivered.  It might be worth you asking Sports Interactive if they plan to make real use of multiple cores and Ram anytime soon.  Extremely frustrating.  

Link to post
Share on other sites

with my old pc i had 16gb of ram and never had any problems. however, it is performing similarly to a laptop i7 running at 2.3 ghz. It should be the same, if not better ipc wise and is running at a much higher clock speed(3.8). So i have concluded that its ram related

Link to post
Share on other sites

39 minutes ago, will221996 said:

with my old pc i had 16gb of ram and never had any problems. however, it is performing similarly to a laptop i7 running at 2.3 ghz. It should be the same, if not better ipc wise and is running at a much higher clock speed(3.8). So i have concluded that its ram related

Yeah 4GB might be a bit low with some OS overhead on top of whatever memory FM is using, that might be your bottleneck.

 

43 minutes ago, bennytee said:

They promised us a truly 64 bit game this year but clearly haven't delivered.  It might be worth you asking Sports Interactive if they plan to make real use of multiple cores and Ram anytime soon.  Extremely frustrating.  

I think SI I have only just started on the path to optimising FM based on 64bit inclusion in FM17 for the first time, next step will be to make FM truly multi-threaded which will take a lot of time and effort as the code has to be split into different threads and coordinated.

Link to post
Share on other sites

1 hour ago, Powermonger said:

I think SI I have only just started on the path to optimising FM based on 64bit inclusion in FM17 for the first time, next step will be to make FM truly multi-threaded which will take a lot of time and effort as the code has to be split into different threads and coordinated.

I imagine it's very difficult to really split the day to day processing into different threads. So many variables apply that will have an effect on other areas of the game. If those are being processed simultaneously coordinating all of that sounds like hell. I only know the very basics of programming though, maybe there's a way. It'd certainly be nice if FM could really use all the multicore horsepower outside of calculating matches.

Link to post
Share on other sites

Thanks @will221996, I've updated the table now with youre Ryzen results. When you add more ram and retest, I'll update your results.

11 hours ago, Urg said:

I imagine it's very difficult to really split the day to day processing into different threads. So many variables apply that will have an effect on other areas of the game. If those are being processed simultaneously coordinating all of that sounds like hell. I only know the very basics of programming though, maybe there's a way. It'd certainly be nice if FM could really use all the multicore horsepower outside of calculating matches.

Yes I can't imagine it being a easy task, it will be a long process.

Link to post
Share on other sites

CPU: Intel i5-2500K

CPU Frequency: 4200 MHz

RAM: 16GB DDR3 800 MHz CL9

OS: Win 7 64-bit

Storage: SSD

Test A

Time: 5 min 31 sec

Test B

Time: 11 min 23sec

Link to post
Share on other sites

On 3/11/2017 at 14:34, jwchriste said:

5-year old system still hanging in there.  Could be better though:

CPU: Intel 3930K

CPU Frequency: 4200 MHz (OC)

RAM: 32GB DDR3 1866 MHz CL10

OS: Win 10 64-bit

Storage: HDD

Test A

Time: 4 min 00 sec

Test B

Time: 7 min 36 sec

Completely new build:

CPU: Intel 7700K

CPU Frequency: 5000 MHz (OC)

RAM: 32GB DDR4 3733 MHz CL18

OS: Win 10 64-bit

Storage: SSD

Test A

Time: 2 min 22 sec

Test B

Time: 4 min 47 sec

Money well spent. :D

Link to post
Share on other sites

2 minutes ago, bennytee said:

@jwchriste wow, nearly twice as fast!

Impressive overclock there, kudos.

Can't see how this can be bettered to be honest, you've got quite the rig there.

I won the silicon lottery with this one tbh.  I've never had this kind of luck with cpus.  It still has headroom.  I had it stable for a short while at 5.2GHz w/ 4000MHz ram, but the voltages required kind of scared me.

Link to post
Share on other sites

7 minutes ago, jwchriste said:

I won the silicon lottery with this one tbh.  I've never had this kind of luck with cpus.  It still has headroom.  I had it stable for a short while at 5.2GHz w/ 4000MHz ram, but the voltages required kind of scared me.

I'd say as long as you have the thermal headroom go for it!

I lost badly with my 4th gen i5 and have to put mega voltage through it to even hit 4.4ghz, i'm secretly hoping the chip dies so the wife will let me build a new rig :D

Link to post
Share on other sites

4 minutes ago, bennytee said:

I'd say as long as you have the thermal headroom go for it!

I lost badly with my 4th gen i5 and have to put mega voltage through it to even hit 4.4ghz, i'm secretly hoping the chip dies so the wife will let me build a new rig :D

I also de-lidded it and put some quality TIM on it.  Well worth it.  Under various stress testing it dropped anywhere from 15-27C.  I couldn't hit 5GHz due to thermals before that.

Just pump even more voltage in yours until you release the magic smoke!

Link to post
Share on other sites

Just now, jwchriste said:

I also de-lidded it and put some quality TIM on it.  Well worth it.  Under various stress testing it dropped anywhere from 15-27C.  I couldn't hit 5GHz due to thermals before that.

Just pump even more voltage in yours until you release the magic smoke!

Your a brave man de-lidding, not sure I have the guts to be honest although that's one way to get a new PC!

That drop in temperature is massive though, to put it in the context of this benchmarking thread de-lidding (whilst not for the feint of heart) has probably shaved maybe 20 seconds off the test which is massive.

Link to post
Share on other sites

37 minutes ago, bennytee said:

Your a brave man de-lidding, not sure I have the guts to be honest although that's one way to get a new PC!

I wouldn't have done it if I had to use a razor blade.  There's a couple tools on the market that work with socket 1150/1151 Intel parts which make it almost idiot-proof.  It does make a horrific sounding "CRACK!" noise though when the lid pops free.  It stopped my heart for moment.

Link to post
Share on other sites

Just done the benchmarking test on my new laptop and here are the results :thup:

CPU: Intel Core i7 - 6700HQ

CPU Frequency: 2601 MHz

RAM: 8GB 2.60GHz

OS: Win 10 64-bit

Storage: SSD

Test A

Time: 4 min 51 sec

Test B

Time: 9 min 37 sec

Link to post
Share on other sites

i'm back... with more ram. Its 12gb of ddr4 2133(not very fast)

CPU: AMD Ryzen R7 1700X

CPU Frequency: 3.4/3.8 GHz

RAM: 12GB DDR4 dual channel

OS: Win 10 64-bit

Storage: SSD

Test A

Time: 4 min 34sec

Test B

Time: 9 min 20sec

ram usage never went above 5.8

maybe add this as a separate entry or a side note to show how ram effects it

i will also add the results for the bigger faster ram to help figure out how faster ram effects fm

finally i this time, more cpu cores were used in test b than last time. 

Link to post
Share on other sites

 

1 hour ago, will221996 said:

i'm back... with more ram. Its 12gb of ddr4 2133(not very fast)

Thanks for testing again, a good improvement with an extra 8GB of RAM and shaved 1 min and 2 min off the initial results. I've added the new results to the table.

@scott-h22 I've added your results too, thanks for testing.

Link to post
Share on other sites

CPU: Intel i5 6600k

CPU Frequency: 4100 MHz // can definitely over clock further

RAM: 16GB DDR4 3000 MHz 

OS: Win 10 pro 64-bit

Storage: SSD

Test A

Time: 3 min 22 sec

Test B

Time: 6min 33 sec

 

Edit: overclocked to 4600mhz

Test A

Time: 3 min 19 sec

Test B

Time: 6min 26 sec

 

Link to post
Share on other sites

On 24/03/2017 at 22:51, scott-h22 said:

Just done the benchmarking test on my new laptop and here are the results :thup:

CPU: Intel Core i7 - 6700HQ

CPU Frequency: 2601 MHz

RAM: 8GB 2.60GHz

OS: Win 10 64-bit

Storage: SSD

Test A

Time: 4 min 51 sec

Test B

Time: 9 min 37 sec

Same setup here & yet im 1 min slower on A & 3 mins slower on B? Hmmmmm

Link to post
Share on other sites

21 minutes ago, giblets_of_fire said:

Same setup here & yet im 1 min slower on A & 3 mins slower on B? Hmmmmm

Could be a cooling issue. Laptop runs too hot, throttles CPU down and FM runs slower.

Link to post
Share on other sites

27 minutes ago, Urg said:

Could be a cooling issue. Laptop runs too hot, throttles CPU down and FM runs slower.

I thought something similar, so i did a test with the fans at 100% and the temps never exceeded 85c (i think 88c is where it tops out & starts throttling) & it was actually a few seconds slower. 

Its not a biggie, its faster than most, just don't like losing!!

Link to post
Share on other sites

10 hours ago, ukmets said:

Just done a Ryzen benchmark

Thanks @ukmets, I've added your results to the table. I justed wanted to confirm the Test A results are correct as it is 1 min faster than someone elses 1700X for Test A but the Test B results are identical. Should it be 4 min 47 sec?

Thanks @jckc221013jamie I have added your results too.

Link to post
Share on other sites

6 hours ago, Powermonger said:

Thanks @ukmets, I've added your results to the table. I justed wanted to confirm the Test A results are correct as it is 1 min faster than someone elses 1700X for Test A but the Test B results are identical. Should it be 4 min 47 sec?

Thanks @jckc221013jamie I have added your results too.

Nope was definitely 3 minutes

Link to post
Share on other sites

5 hours ago, jckc221013jamie said:

thanx for adding me to the table but test b you have put my time as 3 mins 27 secs it was meant to be 6 mins 44 secs  lol

I've corrected now. I was halfway through updating the table when I got wife hate for not coming to dinner :D

4 hours ago, ukmets said:

Nope was definitely 3 minutes

No problems. Cheers. :)

Link to post
Share on other sites

On ‎26‎/‎03‎/‎2017 at 17:50, scott-h22 said:

Could it be because my laptop is brand new?

Well I decided that my laptop could use a fresh start & did a fresh install of Windows (10). Having installed literally just the NVidia drivers and steam my time has gone from 5:56 to 4:54 on test A so something was definitely slowing it down.

Now I just need to determine if it was some software I had installed or one of the myriad of graphics, mods etc that I have on the game (used a completely vanilla version for 4:54)

At least I know the hardware is kosher........

Link to post
Share on other sites

8 minutes ago, giblets_of_fire said:

Well I decided that my laptop could use a fresh start & did a fresh install of Windows (10). Having installed literally just the NVidia drivers and steam my time has gone from 5:56 to 4:54 on test A so something was definitely slowing it down.

Now I just need to determine if it was some software I had installed or one of the myriad of graphics, mods etc that I have on the game (used a completely vanilla version for 4:54)

At least I know the hardware is kosher........

I think it may be more because of your OS patching and driver installs. Depending how old your laptop is, having endless patching getting installed and drivers being updated tends to slow things down after a while. I know in the past that as soon as I installed the OS from scratch and reinstalled my software, speed dramatically improved again.

Link to post
Share on other sites

9 minutes ago, Powermonger said:

I think it may be more because of your OS patching and driver installs. Depending how old your laptop is, having endless patching getting installed and drivers being updated tends to slow things down after a while. I know in the past that as soon as I installed the OS from scratch and reinstalled my software, speed dramatically improved again.

Laptop was 14 months old, but gets used a lot. Going to repeat the test often while I install different bits & bobs

Link to post
Share on other sites

I'm not a Football Manager player but I post on a forum where the topic of what's the best PC spec (within an individual’s given budget) for a PC that will run varying games comes up quite often, and FM17 is a popular question.  I’ve previously linked people to this thread and/or its predecessors for older incarnations of FM.

Since FM17 was free to play this weekend on steam and I had things to be doing on another PC anyway, I took the opportunity to run the benchmarks a few times with slight changes to see what made the most difference.  I'm sure there are no surprises here for the posters in this topic but thought I'd share anyway in case it was of interest.

CPU: i7 3770k
CPU Frequency: 4600MHz
RAM: 16gb DDR3 2400MHz CL10 (2x8gb)
OS: Win 10 Pro 64-bit
Storage:  SSD (1TB Samsung 840 Evo)

 

I ran the benches first with the full spec as above, then also with:
- CPU frequency locked to 3600MHz
- Hyperthreading disabled
- RAM set to 1600MHz CL9
- The game installed on a relatively slow HDD

Each time I tested something the other settings were of course returned to their full level (it wasn’t a progressive nerf of the whole system).

I got slightly different times each time I ran the bench without changing anything, so I ran each one 5 times with each hardware configuration.  These are the averages:

Variable:

Bench A

Bench B:

(full spec as above)

00:03:21

00:06:46

Hyperthreading Off

00:03:31

00:07:10

CPU @ 3600MHz

00:04:07

00:08:25

RAM @ 1600MHz CL9

00:03:39

00:07:31

FM on HDD

00:03:31

00:08:01

(hh:mm:ss)

The HDD usedfor that test was a 1TB Toshiba MK1059GSM, 2.5" 5400rpm (3gbps)
The RAM was G.Skill Trident F3-2400C10D-16GTX, with timings of 10-12-12-31.  When at 1600MHz it was run at CL9 (9-9-9-27)

Sorry that's not more thorough, it was all I had time for.  Would have been nice to test more CPU frequencies, a WD Black HDD, and a 'simulated i3' / the i7 with 2C/4T, plus other RAM frequencies, and to run ten of each.  But I was only able to prodge at this PC to load up the benches and keep an eye on the timer while doing other stuff.

I did record some info about thread usage using HWinfo but would need time to sort through the logs.

Link to post
Share on other sites

Quite a bit of testing there @Oussebon, thanks for posting your results and contributing. Much appreciated.

Your results in different timings seem to follow my own observations. Memory speed and also drive speed can really shave off some time and it would be interesting to see how your results would've looked if you also have had a M.2 NVMe SSD drive to test on.

I've added your results to the table now. :)

Link to post
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...