Upvote Upvoted 0 Downvote Downvoted
1 2 3 4
I'm writing a proper guide to performance.
posted in Q/A Help
1
#1
-4 Frags +

So one day I got irritated at the amount of horrible performance (game performance not the player performance) guides out there. I thought "I can make a better performance guide than most of these guys" so I started work on a comprehensive performance guide. It's not remotely good enough for feedback yet, but I do want to know: What information should I add to my guide pertaining to performance? What are the misconceptions when it comes to TF2 performance. (I'm not talking about the ones on mastercomfig docs I already have those.)

So one day I got irritated at the amount of horrible performance (game performance not the player performance) guides out there. I thought "I can make a better performance guide than most of these guys" so I started work on a comprehensive performance guide. It's not remotely good enough for feedback yet, but I do want to know: What information should I add to my guide pertaining to performance? What are the misconceptions when it comes to TF2 performance. (I'm not talking about the ones on mastercomfig docs I already have those.)
2
#2
41 Frags +
AimIsADickmastercomfig docs

i'm pretty sure those are comprehensive enough

[quote=AimIsADick]mastercomfig docs [/quote]

i'm pretty sure those are comprehensive enough
3
#3
32 Frags +

if you think you can make a better write up than mastercom without making 80% of it her work i'd like to see

if you think you can make a better write up than mastercom without making 80% of it her work i'd like to see
4
#4
-15 Frags +
carterAimIsADickmastercomfig docs
i'm pretty sure those are comprehensive enough

They get the basics down, but the docs are still missing some information. For instance it is not recommended to use dxlevel 81 on modern computers because it uses an older rendering pipeline that is inefficient. Also most people don't realize that framecount per time (like frames per second) is a terrible metric to use for getting performance. They should be using frametime (or frame pacing? Which* term is it?) instead. Not all of this information is readily available and it tends to be hidden away in an obscure post or comment or something like that.

[quote=carter][quote=AimIsADick]mastercomfig docs [/quote]

i'm pretty sure those are comprehensive enough[/quote]

They get the basics down, but the docs are still missing some information. For instance [url=https://www.teamfortress.tv/post/1016536/mastercomfig-fps-customization-config]it is not recommended to use dxlevel 81[/url] on modern computers because it uses an older rendering pipeline that is inefficient. Also most people don't realize that framecount per time (like frames per second) is a terrible metric to use for getting performance. They should be using frametime (or frame pacing? Which* term is it?) instead. Not all of this information is readily available and it tends to be hidden away in an obscure post or comment or something like that.
5
#5
34 Frags +
AimIsADickframes per second) is a terrible metric to use for getting performance.

Yep Cock

[quote=AimIsADick]frames per second) is a terrible metric to use for getting performance.[/quote]

Yep Cock
6
#6
-15 Frags +
carterAimIsADickframes per second) is a terrible metric to use for getting performance.
Yep Cock

No I'm serious. Frames per second doesn't give enough info for it to be useful, and it's often misleading. (Source 1 and Source 2)

[quote=carter][quote=AimIsADick]frames per second) is a terrible metric to use for getting performance.[/quote]

Yep Cock[/quote]

No I'm serious. Frames per second doesn't give enough info for it to be useful, and it's often misleading. ([url=https://www.ropaku.com/what-is-frame-time-why-is-frame-time-important/]Source 1[/url] and [url=https://www.pcgamer.com/why-minimum-fps-can-be-misleading/]Source 2[/url])
7
#7
25 Frags +
AimIsADickcarterAimIsADickmastercomfig docs
i'm pretty sure those are comprehensive enough

They get the basics down, but the docs are still missing some information. For instance it is not recommended to use dxlevel 81 on modern computers because it uses an older rendering pipeline that is inefficient. Also most people don't realize that framecount per time (like frames per second) is a terrible metric to use for getting performance. They should be using frametime (or frame pacing? What term is it?) instead. Not all of this information is readily available and it tends to be hidden away in an obscure post or comment or something like that.

aren't you the dude that was going around to twitch chats saying you had a secret google doc that would help people's performance if they were nice to you?

[quote=AimIsADick][quote=carter][quote=AimIsADick]mastercomfig docs [/quote]

i'm pretty sure those are comprehensive enough[/quote]

They get the basics down, but the docs are still missing some information. For instance [url=https://www.teamfortress.tv/post/1016536/mastercomfig-fps-customization-config]it is not recommended to use dxlevel 81[/url] on modern computers because it uses an older rendering pipeline that is inefficient. Also most people don't realize that framecount per time (like frames per second) is a terrible metric to use for getting performance. They should be using frametime (or frame pacing? What term is it?) instead. Not all of this information is readily available and it tends to be hidden away in an obscure post or comment or something like that.[/quote]

aren't you the dude that was going around to twitch chats saying you had a secret google doc that would help people's performance if they were nice to you?
8
#8
-7 Frags +

https://media.giphy.com/media/fDzM81OYrNjJC/source.gif

[img]https://media.giphy.com/media/fDzM81OYrNjJC/source.gif[/img]
9
#9
-12 Frags +
PeteAimIsADickcarterAimIsADickmastercomfig docs
i'm pretty sure those are comprehensive enough

They get the basics down, but the docs are still missing some information. For instance it is not recommended to use dxlevel 81 on modern computers because it uses an older rendering pipeline that is inefficient. Also most people don't realize that framecount per time (like frames per second) is a terrible metric to use for getting performance. They should be using frametime (or frame pacing? What term is it?) instead. Not all of this information is readily available and it tends to be hidden away in an obscure post or comment or something like that.

aren't you the dude that was going around to twitch chats saying you had a secret google doc that would help people's performance if they were nice to you?

Uh no. I don't remember ever doing that.

[quote=Pete][quote=AimIsADick][quote=carter][quote=AimIsADick]mastercomfig docs [/quote]

i'm pretty sure those are comprehensive enough[/quote]

They get the basics down, but the docs are still missing some information. For instance [url=https://www.teamfortress.tv/post/1016536/mastercomfig-fps-customization-config]it is not recommended to use dxlevel 81[/url] on modern computers because it uses an older rendering pipeline that is inefficient. Also most people don't realize that framecount per time (like frames per second) is a terrible metric to use for getting performance. They should be using frametime (or frame pacing? What term is it?) instead. Not all of this information is readily available and it tends to be hidden away in an obscure post or comment or something like that.[/quote]

aren't you the dude that was going around to twitch chats saying you had a secret google doc that would help people's performance if they were nice to you?[/quote]

Uh no. I don't remember ever doing that.
10
#10
24 Frags +

also as everyone already told you
you wouldn't have to care about half this shit if you just saved some money and got an actual pc

also as everyone already told you
you wouldn't have to care about half this shit if you just saved some money and got an actual pc
11
#11
19 Frags +
AimIsADick (like frames per second) is a terrible metric to use for getting performance

https://youtu.be/pXAQNGUcTjs

[quote=AimIsADick] (like frames per second) is a terrible metric to use for getting performance [/quote]

[youtube]https://youtu.be/pXAQNGUcTjs[/youtube]
12
#12
-8 Frags +
Brimstonealso as everyone already told you
you wouldn't have to care about half this shit if you just saved some money and got an actual pc

No I'm talking about that (those frametime spikes seem to be exaggerated btw. Just playing in some TF2 lobbies with cl_showfps 2 shows me that's it's generally pretty low like around 16.7ms, so it's likely the software I used for testing). I just got tired of those shitty performance guides online (like that one made by Delfy and that one other youtuber that said mastercomfig made his game run worse).

[quote=Brimstone]also as everyone already told you
you wouldn't have to care about half this shit if you just saved some money and got an actual pc[/quote]

No I'm talking about that (those frametime spikes seem to be exaggerated btw. Just playing in some TF2 lobbies with cl_showfps 2 shows me that's it's generally pretty low like around 16.7ms, so it's likely the software I used for testing). I just got tired of those shitty performance guides online (like that one made by Delfy and that one other youtuber that said mastercomfig made his game run worse).
13
#13
8 Frags +

Aim Is A Dick

Aim Is A Dick
14
#14
Stream Highlights
23 Frags +

Have you checked this one out?

Have you checked [url=https://pastebin.com/ZC5h7Bec]this one[/url] out?
15
#15
6 Frags +

bullying aim on a tuesday

bullying aim on a tuesday
16
#16
16 Frags +

https://media.discordapp.net/attachments/779486942370594827/819017906713002025/unknown.png

this is from the first source that you linked in #6

"frametime" and frames per second are the same fucking thing lol

you cannot seriously think you know more than mastercoms about tf2

AimIsADickI just got tired of those shitty performance guides online (like that one made by Delfy and that one other youtuber that said mastercomfig made his game run worse).

If you're referencing this, I can't find a single thing in this video that is wrong or shitty. it's geared at pubbers. it's not made with disabling literally every graphical setting in the game in mind lol

[img]https://media.discordapp.net/attachments/779486942370594827/819017906713002025/unknown.png[/img]

this is from the first source that you linked in #6

"frametime" and frames per second are the same fucking thing lol

you cannot seriously think you know more than mastercoms about tf2

[quote=AimIsADick]I just got tired of those shitty performance guides online (like that one made by Delfy and that one other youtuber that said mastercomfig made his game run worse).[/quote]

If you're referencing [url=https://www.youtube.com/watch?v=udPpDmCuCV0]this[/url], I can't find a single thing in this video that is wrong or shitty. it's geared at pubbers. it's not made with disabling literally every graphical setting in the game in mind lol
17
#17
-23 Frags +

I'm starting to think this project was a bad idea.

carterhttps://media.discordapp.net/attachments/779486942370594827/819017906713002025/unknown.png

this is from the first source that you linked in #6

"frametime" and frames per second are the same fucking thing lol

They're the same metric type, not actually the same metric. The difference between the two is in how the information is represented. Frametime represents the time it takes to render the frame, while frames per second represents the amount of frames generated in a second. What I'm saying is that frametimes are generally better than frames per second as a benchmark metric, because it gives more information.

carteryou cannot seriously think you know more than mastercoms about tf2

Well yeah I don't. I know fuck all about Source and I'd rather not try to bother with it.

carterAimIsADickI just got tired of those shitty performance guides online (like that one made by Delfy and that one other youtuber that said mastercomfig made his game run worse).
If you're referencing this, I can't find a single thing in this video that is wrong or shitty. it's geared at pubbers. it's not made with disabling literally every graphical setting in the game in mind lol
  1. This video presupposes that more FPS = better performance even though that isn't always the case.
  2. dxlevel 95 is only available on xbox 360. On PC the dxlevel will be maxed to 90.
  3. mat_queue_mode should be left to -1. Let the game figure out when to use queue mode instead of forcing it on or off.
  4. Any evidence for these claims made in this video?
I'm starting to think this project was a bad idea.

[quote=carter][img]https://media.discordapp.net/attachments/779486942370594827/819017906713002025/unknown.png[/img]

this is from the first source that you linked in #6

"frametime" and frames per second are the same fucking thing lol[/quote]

They're the same metric [i]type[/i], not actually the same metric. The difference between the two is in how the information is represented. Frametime represents the [i]time it takes to render the frame[/i], while frames per second represents the [i]amount of frames generated in a second[/i]. What I'm saying is that frametimes are generally better than frames per second as a benchmark metric, because it gives more information.

[quote=carter]you cannot seriously think you know more than mastercoms about tf2[/quote]

Well yeah I don't. I know fuck all about Source and I'd rather not try to bother with it.

[quote=carter][quote=AimIsADick]I just got tired of those shitty performance guides online (like that one made by Delfy and that one other youtuber that said mastercomfig made his game run worse).[/quote]

If you're referencing [url=https://www.youtube.com/watch?v=udPpDmCuCV0]this[/url], I can't find a single thing in this video that is wrong or shitty. it's geared at pubbers. it's not made with disabling literally every graphical setting in the game in mind lol[/quote]

[olist]
[*] This video presupposes that more FPS = better performance even though that isn't always the case.
[*] dxlevel 95 is only available on xbox 360. On PC the dxlevel will be maxed to 90.
[*] mat_queue_mode should be left to -1. Let the game figure out when to use queue mode instead of forcing it on or off.
[*] Any evidence for these claims made in this video?
[/olist]
18
#18
25 Frags +

There are two incorrect things in the Delfy video that I spotted: r_lod 2 instead of r_rootlod 2 and the recommendation to use -threads -high in launch options (granted, I skimmed through it).

Anyways, the mastercomfig docs are open source! So you can improve them however you'd like.

AimIsADickThey're the same metric type, not actually the same metric. The difference between the two is in how the information is represented. Frametime represents the time it takes to render the frame, while frames per second represents the amount of frames generated in a second. What I'm saying is that frametimes are generally better than frames per second as a benchmark metric, because it gives more information.

How do you think a game engine or any benchmarking tool measures FPS? It uses frametime by measuring the time taken between the start and end of the frame, then takes the reciprocal of that (divides 1/frametime).

AimIsADickdxlevel 95 is only available on xbox 360. On PC the dxlevel will be maxed to 90.

You may be thinking of dxlevel 98.

There are two incorrect things in the Delfy video that I spotted: r_lod 2 instead of r_rootlod 2 and the recommendation to use -threads -high in launch options (granted, I skimmed through it).

Anyways, the mastercomfig docs are open source! So you can improve them however you'd like.

[quote=AimIsADick]
They're the same metric [i]type[/i], not actually the same metric. The difference between the two is in how the information is represented. Frametime represents the [i]time it takes to render the frame[/i], while frames per second represents the [i]amount of frames generated in a second[/i]. What I'm saying is that frametimes are generally better than frames per second as a benchmark metric, because it gives more information.
[/quote]

How do you think a game engine or any benchmarking tool measures FPS? It uses frametime by measuring the time taken between the start and end of the frame, then takes the reciprocal of that (divides 1/frametime).

[quote=AimIsADick]
dxlevel 95 is only available on xbox 360. On PC the dxlevel will be maxed to 90.[/quote]
You may be thinking of dxlevel 98.
19
#19
-12 Frags +
mastercomsThere are two incorrect things in the Delfy video that I spotted: r_lod 2 instead of r_rootlod 2 and the recommendation to use -threads -high in launch options (granted, I skimmed through it).

Anyways, the mastercomfig docs are open source! So you can improve them however you'd like.

Yeah ik. I was just thinking of tips that weren't necessarily related to your config (like if neccessary increasing the page file size on windows, cleaning up bloatware, etc.).

mastercomsAimIsADickThey're the same metric type, not actually the same metric. The difference between the two is in how the information is represented. Frametime represents the time it takes to render the frame, while frames per second represents the amount of frames generated in a second. What I'm saying is that frametimes are generally better than frames per second as a benchmark metric, because it gives more information.
How do you think a game engine or any benchmarking tool measures FPS? It uses frametime by measuring the time taken between the start and end of the frame, then takes the reciprocal of that (divides 1/frametime).

I must have been thinking of something else when I thought that they weren't the same metric, but yeah they are the same metric. (Sorry about that!)

mastercomsAimIsADickdxlevel 95 is only available on xbox 360. On PC the dxlevel will be maxed to 90.You may be thinking of dxlevel 98.

Yeah I was thinking of dxlevel 98 for some reason.

[quote=mastercoms]There are two incorrect things in the Delfy video that I spotted: r_lod 2 instead of r_rootlod 2 and the recommendation to use -threads -high in launch options (granted, I skimmed through it).

Anyways, the mastercomfig docs are open source! So you can improve them however you'd like.[/quote]

Yeah ik. I was just thinking of tips that weren't necessarily related to your config (like [i]if neccessary[/i] increasing the page file size on windows, cleaning up bloatware, etc.).

[quote=mastercoms][quote=AimIsADick]
They're the same metric [i]type[/i], not actually the same metric. The difference between the two is in how the information is represented. Frametime represents the [i]time it takes to render the frame[/i], while frames per second represents the [i]amount of frames generated in a second[/i]. What I'm saying is that frametimes are generally better than frames per second as a benchmark metric, because it gives more information.
[/quote]

How do you think a game engine or any benchmarking tool measures FPS? It uses frametime by measuring the time taken between the start and end of the frame, then takes the reciprocal of that (divides 1/frametime).[/quote]

I must have been thinking of something else when I thought that they weren't the same metric, but yeah they are the same metric. (Sorry about that!)

[quote=mastercoms][quote=AimIsADick]
dxlevel 95 is only available on xbox 360. On PC the dxlevel will be maxed to 90.[/quote]
You may be thinking of dxlevel 98.[/quote]

Yeah I was thinking of dxlevel 98 for some reason.
20
#20
11 Frags +

You may consider collaborate with this gentleman https://www.teamfortress.tv/56288/fps-config-based-on-research-and-not-vpk

You may consider collaborate with this gentleman https://www.teamfortress.tv/56288/fps-config-based-on-research-and-not-vpk
21
#21
16 Frags +

i cannot wait for AimIsADick to start playing 6s

i cannot wait for AimIsADick to start playing 6s
22
#22
21 Frags +
charisi cannot wait for AimIsADick to start playing 6s

we may need to start a gofundme to buy him a 50 dollar discrete gpu from like 2010 to make that happen

[quote=charis]i cannot wait for AimIsADick to start playing 6s[/quote]

we may need to start a gofundme to buy him a 50 dollar discrete gpu from like 2010 to make that happen
23
#23
3 Frags +
AimIsADickframes per second) is a terrible metric to use for getting performance.

i think you just mean that sometimes it's better to have a stable framerate rather than an unstable but high-ish one. then the frametimes are also stable, because it really is kinda the same thing (so the experience is far more enjoyable)

[quote=AimIsADick]frames per second) is a terrible metric to use for getting performance.[/quote]

i think you just mean that sometimes it's better to have a stable framerate rather than an unstable but high-ish one. then the frametimes are also stable, because it really is kinda the same thing (so the experience is far more enjoyable)
24
#24
21 Frags +
AimIsADickThey should be using frametime (or frame pacing? What term is it?) instead.

You know its going to be a good config when OP proposes to not use frames per second and instead suggests using something he forgot the name of...

[quote=AimIsADick]They should be using frametime (or frame pacing? What term is it?) instead.[/quote]

You know its going to be a good config when OP proposes to not use frames per second and instead suggests using something he forgot the name of...
25
#25
2 Frags +

Maybe I misunderstand what frametime is, but isn't frame time more granular than FPS...? B/c you get a time for every frame and can see very detailed oscillations that FPS hides. (like one frame takes 10ms to render and the next spikes to 150ms for some reason.) FPS gives aggregate info over the second so it loses some of that detail. The mean or total frametime over the second is the same as FPS, but you can't recover individual frametimes from the FPS... So, they're not the same metric..., right?
I need someone to explain it to me. ;-;

Maybe I misunderstand what frametime is, but isn't frame time more granular than FPS...? B/c you get a time for every frame and can see very detailed oscillations that FPS hides. (like one frame takes 10ms to render and the next spikes to 150ms for some reason.) FPS gives aggregate info over the second so it loses some of that detail. The mean or total frametime over the second is the same as FPS, but you can't recover individual frametimes from the FPS... So, they're not the same metric..., right?
I need someone to explain it to me. ;-;
26
#26
-10 Frags +
vaniAimIsADickThey should be using frametime (or frame pacing? What term is it?) instead.
You know its going to be a good config when OP proposes to not use frames per second and instead suggests using something he forgot the name of...

I made a typo there. I meant which term I should use. I was unsure whether to use frametime or frame pacing.

Bob_MarleyMaybe I misunderstand what frametime is, but isn't frame time more granular than FPS...? B/c you get a time for every frame and can see very detailed oscillations that FPS hides. (like one frame takes 10ms to render and the next spikes to 150ms for some reason.) FPS gives aggregate info over the second so it loses some of that detail. The mean or total frametime over the second is the same as FPS, but you can't recover individual frametimes from the FPS... So, they're not the same metric..., right?
I need someone to explain it to me. ;-;

They are the same metric but yes frametime is more granular than frames per second. This is what I was trying to explain to these people.

[quote=vani][quote=AimIsADick]They should be using frametime (or frame pacing? What term is it?) instead.[/quote]

You know its going to be a good config when OP proposes to not use frames per second and instead suggests using something he forgot the name of...[/quote]

I made a typo there. I meant [i]which[/i] term I should use. I was unsure whether to use frametime or frame pacing.

[quote=Bob_Marley]Maybe I misunderstand what frametime is, but isn't frame time more granular than FPS...? B/c you get a time for every frame and can see very detailed oscillations that FPS hides. (like one frame takes 10ms to render and the next spikes to 150ms for some reason.) FPS gives aggregate info over the second so it loses some of that detail. The mean or total frametime over the second is the same as FPS, but you can't recover individual frametimes from the FPS... So, they're not the same metric..., right?
I need someone to explain it to me. ;-;[/quote]

They are the same metric but yes frametime is more granular than frames per second. This is what I was trying to explain to these people.
27
#27
8 Frags +
AimIsADick "I can make a better performance guide than most of these guys"

ok, do it then, instead of making this thread.

[quote=AimIsADick] "I can make a better performance guide than most of these guys"[/quote]
ok, do it then, instead of making this thread.
28
#28
5 Frags +
AimIsADick (game performance not the player performance)

yeah that's smart, something tells me you wouldn't really be able to talk on that either

[quote=AimIsADick] (game performance not the player performance)[/quote] [url=https://i.imgur.com/4uhx1WD.png]yeah that's smart, something tells me you wouldn't really be able to talk on that either[/url]
29
#29
-8 Frags +
kbtoAimIsADick (game performance not the player performance) yeah that's smart, something tells me you wouldn't really be able to talk on that either

Ah yes an ad hominem attack.

(I think posting about the guide here was a mistake after all.)

[quote=kbto][quote=AimIsADick] (game performance not the player performance)[/quote] [url=https://i.imgur.com/4uhx1WD.png]yeah that's smart, something tells me you wouldn't really be able to talk on that either[/url][/quote]

Ah yes an ad hominem attack.

(I think posting about the guide here was a mistake after all.)
30
#30
7 Frags +
AimIsADickkbtoAimIsADick (game performance not the player performance) yeah that's smart, something tells me you wouldn't really be able to talk on that either
Ah yes an ad hominem attack.

(I think posting about the guide here was a mistake after all.)

how tf is any of this ad hominem im saying youre unqualified to talk since your most played class is 150 hours on spy?

[quote=AimIsADick][quote=kbto][quote=AimIsADick] (game performance not the player performance)[/quote] [url=https://i.imgur.com/4uhx1WD.png]yeah that's smart, something tells me you wouldn't really be able to talk on that either[/url][/quote]

Ah yes an ad hominem attack.

(I think posting about the guide here was a mistake after all.)[/quote]
how tf is any of this ad hominem im saying youre unqualified to talk since your most played class is 150 hours on spy?
1 2 3 4
Please sign in through STEAM to post a comment.