FM20 Performance Benchmarking Thread

[roykela]

Either way is fine with me.

[Smurf]

@Brother BenIt's your call. Maybe highlight it's Beta testing. And perhaps when the full verson is released we can have the thread split and the Beta one locked? @FrazT@EdL Any ideas?

[EdL]

At this point I'd just wait for the full release, ideally you want everyone testing with the same FM version as much as possible

[Brother Ben]

Okay i'll wait.  It's less than a week away at this point anyway

[Brother Ben]

Post 2 and spreadsheet updated

@kevhamster i've left off your second benchmark as there's something clearly wrong there.  Blocked vents maybe?  Fan issue?

[kevhamster]

3 minutes ago, Brother Ben said:

Post 2 and spreadsheet updated

@kevhamster i've left off your second benchmark as there's something clearly wrong there.  Blocked vents maybe?  Fan issue?

I'll try and open the thing up and see if there's any dust issues or anything.

It seems to be in line with the 5th gen i5's as things stand!

 

[Smurf]

11 minutes ago, Brother Ben said:

Post 2 and spreadsheet updated

@kevhamster i've left off your second benchmark as there's something clearly wrong there.  Blocked vents maybe?  Fan issue?

On the combined sheet - A and B seem to be missing, or I can't work out how to view them...???

[Brother Ben]

13 minutes ago, Smurf said:

On the combined sheet - A and B seem to be missing, or I can't work out how to view them...???

I'd hidden them to do the screenshots, should be fine now

[Brother Ben]

Okay so here's a preview of what the benchmarks will look like for FM21

I've standardised the benchmarking dates after a fair bit of confusion

 

[kevhamster]

What I have discovered today.... HP really don't like people being able to pop the back off of their laptops!

 

[Brother Ben]

3 minutes ago, kevhamster said:

What I have discovered today.... HP really don't like people being able to pop the back off of their laptops!

 

Tell me about it, i've got a 17 inch HP Envy and its the worst laptop i've ever had the misfortune to open up.  I actually wrecked the hinges too

[kevhamster]

Just now, Brother Ben said:

Tell me about it, i've got a 17 inch HP Envy and its the worst laptop i've ever had the misfortune to open up.  I actually wrecked the hinges too

I've popped all the bits round the outside, so it's partially there.  Seems to be stuck somewhere in the middle!

 

[kevhamster]

Finally managed to open it up after worrying I'd wreck the back panel.  Was a tiny amount of dust, but nothing which I'd consider should be causing an issue.  Still, I will try the benchmark again later and see how it compares as I did give the fan a quick dust over just in case.

 

[kevhamster]

I just re-ran benchmark A on the laptop, and got a result of 2 minutes and 51 seconds, which is a reasonably significant difference to the result from yesterday (3 minutes 28 seconds).  If my maths serves me correct, that an improvement of 18% after a little dust, as that's all I did.... didn't even look like much dust.

I will run the others later and see how those pan out.  Will be especially interesting as the bottom of the laptop got quite hot for those more intensive benchmark tests.

 

[craiigman]

46 minutes ago, Brother Ben said:

Okay so here's a preview of what the benchmarks will look like for FM21

I've standardised the benchmarking dates after a fair bit of confusion

 

Happy to put the M1 through its paces again once the saves are available.

[Brother Ben]

40 minutes ago, kevhamster said:

I just re-ran benchmark A on the laptop, and got a result of 2 minutes and 51 seconds, which is a reasonably significant difference to the result from yesterday (3 minutes 28 seconds).  If my maths serves me correct, that an improvement of 18% after a little dust, as that's all I did.... didn't even look like much dust.

I will run the others later and see how those pan out.  Will be especially interesting as the bottom of the laptop got quite hot for those more intensive benchmark tests.

 

Still seems a bit slow.  Is there any other programs hogging memory or processor power?

[kevhamster]

1 minute ago, Brother Ben said:

Still seems a bit slow.  Is there any other programs hogging memory or processor power?

Nothing obvious in task manager.

[Brother Ben]

32 minutes ago, kevhamster said:

Nothing obvious in task manager.

Strange one then

You can use this program to monitor whats going on https://www.cpuid.com/downloads/hwmonitor/hwmonitor_1.42.exe

it'll show you your clockspeeds and you can see if they throttle right down.  It won't solve your problem but at least you'll know there is one if it clocks down really quick

it also shows temps so you can see if thats an issue.

Ideally it'll be a software issue and although a pain in the arse at least you know you can just re-install windows and have it like new

[kevhamster]

10 minutes ago, Brother Ben said:

Strange one then

You can use this program to monitor whats going on https://www.cpuid.com/downloads/hwmonitor/hwmonitor_1.42.exe

it'll show you your clockspeeds and you can see if they throttle right down.  It won't solve your problem but at least you'll know there is one if it clocks down really quick

it also shows temps so you can see if thats an issue.

Ideally it'll be a software issue and although a pain in the arse at least you know you can just re-install windows and have it like new

Yeah, that would be a pain in the arse if I need to reinstall Windows.  I use the laptop as my main computer when it comes to my music production and there are so many different programs and plugins to install.... took me three whole evenings last time when I first got it lol (my desktop, although it's a lot stronger, is also a fair bit noisier due to the rear fan being in constant operation!).

But thanks for the program link though, I'll install that and see what's going on with it when I try the second benchmark later.

 

[EdL]

1 hour ago, kevhamster said:

Yeah, that would be a pain in the arse if I need to reinstall Windows.  I use the laptop as my main computer when it comes to my music production and there are so many different programs and plugins to install.... took me three whole evenings last time when I first got it lol (my desktop, although it's a lot stronger, is also a fair bit noisier due to the rear fan being in constant operation!).

But thanks for the program link though, I'll install that and see what's going on with it when I try the second benchmark later.

 

Even plain old task manager will show you the current clock frequency

[kevhamster]

@Brother Ben @EdL

Now things get a little interesting.

Ran benchmark C again out of curiosity.  This time, it was worse, with 21 minutes and 23 seconds (so approx 30 seconds longer).

Apologies for poor quality pics, but I took them with my phone - didn't want to do screenshots and put extra load on the computer.

This was fairly typical for large parts of the test:

So, that all looks OK.  However, once the game hit 31st August. this was more typical:

All cores hitting 100% usage, clearly being throttled down to 1.60GHz (before 31st August, the brief times I did get 100% on all cores, they remained at 2.00GHz).

Despite that, the underside of the laptop felt significantly less hot than it did yesterday.

[Brother Ben]

@kevhamster

Gotta say the temps don't look too bad, pretty good in fact.  Not sure why its throttling to less than half its top speed.

Without wanting to sound patronising have you set the power plan to high performance and do you have the latest drivers?

May sound a bit mad but try the benchmark with the power not plugged in, just on battery power and see if there's a difference

also if all else fails you could try putting it on power saving mode, it sounds counter intuitive but it may well hold a nice middle ground of clock speed so it won't throttle

Edited November 19, 2020 by Brother Ben

[kevhamster]

15 minutes ago, Brother Ben said:

@kevhamster

Gotta say the temps don't look too bad, pretty good in fact.  Not sure why its throttling to less than half its top speed.

Without wanting to sound patronising have you set the power plan to high performance and do you have the latest drivers?

May sound a bit mad but try the benchmark with the power not plugged in, just on battery power and see if there's a difference

also if all else fails you could try putting it on power saving mode, it sounds counter intuitive but it may well hold a nice middle ground of clock speed so it won't throttle

It is plugged in (I always have it plugged in) and I always set it to high performance whenever I get a new laptop! 

I may try that and see how it goes. Thanks! 

[Brother Ben]

1 minute ago, kevhamster said:

It is plugged in (I always have it plugged in) and I always set it to high performance whenever I get a new laptop! 

I may try that and see how it goes. Thanks! 

it sounds odd but my laptop works better when it isn't plugged in.

I googled your Ryzen and HP and there are quite a few complaining of the same issue, its a pain, I definitely won't be getting another HP when mine packs in

[jonpt]

On 15/11/2020 at 09:55, Daveincid said:

Awesome! I'm kinda jealous with my 3950x. Do you have any possiblility to OC the RAM to 4000MHz?  Or activate PBO? Just to see if there is still room for improvement?

 

Cheers

 

Daveincid

I’ve had a little play around. Ram wise I’m limited, I think a BIOS revision may help here as I’m on a beta bios.

for PBO it does make a bit of difference. The cpu will boost pass 5ghz on single core and 4.6 on all cores. 
 

i’ve tried a Static overclock but the power consumption goes up massively and I can only get 4.4 stable. 

[Daveincid]

vor 1 Minute schrieb jonpt:

I’ve had a little play around. Ram wise I’m limited, I think a BIOS revision may help here as I’m on a beta bios.

for PBO it does make a bit of difference. The cpu will boost pass 5ghz on single core and 4.6 on all cores. 
 

i’ve tried a Static overclock but the power consumption goes up massively and I can only get 4.4 stable. 

Thanks for taking the time. Sounds solid! PBO doesn't worked that well on my 3950x. Have your runned the Test with PBO activated again, to show the difference in fm?

I was able to get some pretty decent results with OC and undervolting in terms of Power-consumption/heat. 

[Kruj]

Am 19.11.2020 um 14:28 schrieb Brother Ben:

Okay so here's a preview of what the benchmarks will look like for FM21

I've standardised the benchmarking dates after a fair bit of confusion

 

Wow this is great! Finally the benchmarks simulate the entire database with all players! I'm looking forward to the results! Definetely makes my choice for a new CPU easier because I always wanted to play the FM with the entire DB.

Edited November 21, 2020 by Kruj

[kevhamster]

29 minutes ago, Kruj said:

Wow this is great! Finally the benchmarks simulate the entire database with all players! I'm looking forward to the results! Definetely makes my choice for a new CPU easier because I always wanted to play the FM with the entire DB.

The benchmarks for FM20 also used full databases (specifically benchmark C). 

[Brother Ben]

3 hours ago, kevhamster said:

The benchmarks for FM20 also used full databases (specifically benchmark C). 

Actually I only used "large" mainly because it was what had been done before but this time I thought why not?

Edited November 21, 2020 by Brother Ben

[kevhamster]

8 minutes ago, Brother Ben said:

Actually I only used "large" mainly because it was what had been done before but this time I thought why not?

Ah OK. My misunderstanding! 

[Brother Ben]

1 minute ago, kevhamster said:

Ah OK. My misunderstanding! 

I'll be honest Kev its not a definite yet it would take the player count from 162,590 to 392, 590 which is huge.

I'll test the difference myself before I decide

[kevhamster]

2 minutes ago, Brother Ben said:

I'll be honest Kev its not a definite yet it would take the player count from 162,590 to 392, 590 which is huge.

I'll test the difference myself before I decide

Don't SI recommend a maximum of 120,000?  Or was that just older editions of FM? 

[Brother Ben]

1 minute ago, kevhamster said:

Don't SI recommend a maximum of 120,000?  Or was that just older editions of FM? 

Yeah for long term stability but for our purposes here it will (should) be fine

[Kruj]

I think it's great. There are in fact some crazy guys who aren't happy enough with a "large" database and thus want to have everything I'm one of these and my buying decision for a new CPU depends on that.

 

Maybe we can add a "Benchmark D" which reflects the former Benchmark C with about 150.000 players?!

[Daveincid]

Gerade eben schrieb Kruj:

I think it's great. There are in fact some crazy guys who aren't happy enough with a "large" database and thus want to have everything I'm one of these and my buying decision for a new CPU depends on that.

 

Maybe we can add a "Benchmark D" which reflects the former Benchmark C with about 150.000 players?!

Let me introduce myself

I would really appreciate a Benchmark "D" with literally EVERYTHING. 

[Brother Ben]

3 hours ago, Daveincid said:

Let me introduce myself

I would really appreciate a Benchmark "D" with literally EVERYTHING. 

I understand what you mean but its quite an ask getting people to do 3 benchmarks never mind 4

Plus i'm betting if you do everything including full detail it'll only be feasable for the most top end of desktops, I can't imagine any laptops being able to cope

I'll run a few tests tonight but I think the tests may well end up being the same as last year but with the holiday times standardised 

Edited November 23, 2020 by Brother Ben

[Daveincid]

vor 5 Minuten schrieb Brother Ben:

I understand what you mean but its quite an ask getting people to do 3 benchmarks never mind 4

Plus i'm betting if you do everything including full detail it'll only be feasable for the most top end of desktops, I can't imagine any laptops being able to cope

I understand your point, I bet that too, but I don't know it, that's why I really love your benchmarks, you get facts. But there are people who play this way, and as Kruj said, it really do have an influence for some future decisions what to buy.  Or add it as a "feel free to break your system-option"? I would really looking forward to see results in this area.

Cheers

Daveincid

Edited November 23, 2020 by Daveincid
grammar

[Brother Ben]

2 minutes ago, Daveincid said:

I understand your point, I bet that too, but I don't know it, that's why I really love your benchmarks, you get facts. But there are people who play this way, and as Kruj said, it really do have an influence for some future decisions what to buy.  Or add it as a "feel free to break your system-option"? I would really looking forward to see results in this area.

Cheers

Daveincid

Maybe I could make that the optional Benchmark D with a warning disclaimer 

[Daveincid]

Gerade eben schrieb Brother Ben:

Maybe I could make that the optional Benchmark D with a warning disclaimer 

Would be really nice! But it's really up to you!

[Brother Ben]

We'll see, i'm testing it now but all leagues and all players is already taking an AGE to setup!

The trouble this year as well is that you have to do the test in the second season because some leagues don't start until then.  The latest startdate is May 2021 (Iceland) so I have to holiday from early May to late August which is no mean feat in itself  

[Daveincid]

Gerade eben schrieb Brother Ben:

We'll see, i'm testing it now but all leagues and all players is already taking an AGE to setup!

The trouble this year as well is that you have to do the test in the second season because some leagues don't start until then.  The latest startdate is May 2021 (Iceland) so I have to holiday from early May to late August which is no mean feat in itself  

Ouch! I haven't thought about that  No pressure at all, for sure

[Brother Ben]

Just now, Daveincid said:

Ouch! I haven't thought about that  No pressure at all, for sure

I estimate based on last years benchmark B that it'll take me at least 2-3 hours to get to August.  Does anyone know when the actual full game will be released?  It would obviously be easier for me to do this over night 

[kevhamster]

4 minutes ago, Brother Ben said:

I estimate based on last years benchmark B that it'll take me at least 2-3 hours to get to August.  Does anyone know when the actual full game will be released?  It would obviously be easier for me to do this over night 

All I know is that it's tomorrow at some point.  I assume SI will make an announcement when the retail launch has happened.

[Joey Numbaz]

5 hours ago, Brother Ben said:

Plus i'm betting if you do everything including full detail it'll only be feasable for the most top end of desktops, I can't imagine any laptops being able to cope

I have run my FM20 save on a 4 year old MacBook Pro with 410,000 players and every league including youth on full detail. Saturdays can be slow, but it definitely works. I am in 2026-27 right now in terms of long-term issues.

[marioo]

When will version 2021 be available?

[Brother Ben]

13 minutes ago, marioo said:

When will version 2021 be available?

Ask and you shall receive 

 

 

[maaartio]

Type: Laptop

Model: PC Desktop

CPU Model: Ryzen 5900 x

CPU Base Frequency: stock

CPU Turbo Frequency: stock

RAM: 16GB

RAM Clockspeed: 2128 Mhz  or 3200 Mhz

GPU: radeon rx 480

Storage Type: SSD 980 pro Samsung

 

Benchmark A: 44 s (2128 Mhz RAM);                                43 s (3200 RAM)

Benchmark B: 4 min 35 sec (2128 Mhz RAM)                  4 min 12 sec (3200 RAM)

Benchmark C: 3 min 38 sec (2128 Mhz RAM)                  3 min 22 sec (3200 RAM)

Edited December 1, 2020 by maaartio

[Cecin]

3700x

16gb - 3400mhz Cl 18

HD 1tb - WD 7200rpm

Rx 470 XFX

No Overclock, only Memory....

A - 55.5 seg
B - 5 min 50 seg
C - 4 min 54 seg

 

 

[Slowpoke83]

Just saw this thread while searching for a new PC to play FM.

I got an old office laptop and my speed is very bad...

 

Type: Laptop

Model: Lenovo G710

CPU Model: i5 - 4210M

CPU Base Frequency: 2.5GHz

CPU Turbo Frequency: 3.20 GHz

RAM: 8GB DDR3

RAM Clockspeed: 1600Mhz

GPU: Intel® HD-Grafik 4600

Storage Type: SSD

 

I only did Benchmark A: 3 min 59 sec

[Kruj]

Type: Desktop PC

Model: Custom

CPU Model: AMD Ryzen 5900X

CPU Base Frequency: 3.7GHz

CPU Turbo Frequency: 4.8 GHz

RAM: 64GB

RAM Clockspeed: 3600Mhz

GPU: AMD Radeon RX 5600XT

Storage Type: NVMe-SSD

 

Benchmark C: 3:31

 

Too bad I don't have FM21 yet, I'd like to do the Benchmark D test from the FM21 Benchmark thread