DOOM.EXE -TIMEDEMO DEMO3

starts the benchmark. To have comparable results with that benchmark pack I shard, make sure sound and music is disabled and screensize set to full screen (no HUD).

To workout the FPS use this formular:

74690 / ticks

Hi Phil,

That's great thanks!

I manage to load it up ok with the extension and the game runs a bit slower than if just normal DOOM.EXE. Though when I exit the game back to dos via the menu... no realtics are listed.

How should I exit the game? I tried dying in game, but the character just stayed on the floor. 🤣

Thanks for your help... almost there

It should launch, then the benchmark should commence and when the character dies it should drop to DOS and output the results!

Don't tough anything while it's running 😀

It's a timedemo, so it plays back an existing recording, you don't play yourself.

Excellent, thank you.

I was going to say... your shooting was a bit rusty there! 🤣

It worked out to be 2134 gametics in 8958 realtics which gives I believe, 8.34 fps.

When playing the game without the demo and in full screen... the game is very much fluid and playable with no lag as there is in the time demo.

I'll remember this now for future benching, thanks

The above result was from the Compaq Systempro in Dos (only 1 386DX40 CPU) and can be added to the rest of the results from here; -

The Systempro Project; The Build

Trying to find more Windows NT benchmarks as it is under this OS that both CPU's run.

Thanks

I wouldn't recommend to bench the way "PhilsComputerLab" is suggesting, using and constantly promoting.

Almost everyone plays Doom with the HUD enabled. So the realistic size is fullscreen -1.
In more than 20 years, I never met anyone who played Doom without the HUD.
So one has to bench Doom with fullscreen -1 (HUD enabled) to obtain real world benchmark results which are comarable to all the available results on ancient websites and newsgroups. These results are a bit higher (more fps), because the CPU has to render a slightly less complex scene.

on one hand side i ask here: how much of a difference does it really make? results please?
on the other: since when have benchmarks anything to do with real life performance? no music, no input to process (sometimes its a whole separate main loop in the program), benchmarks run "at any speed", while in real life you always look for settings that are playable in effect putting a different kind of load on the machine?
im just sayan :>

I agree. Benchmarking (from a PC perspective) is purely a method of comparing the performance of different systems against each other, using the same parameters.

It is not meant to necessarily achieve or produce real world performance (since the are so many other factors that can influence real world performance).

Simply one point: If there are hundereds of results available, using the method with the HUD enabled, whats the point to promote a different method that is not comparable to all the existing work of all the folks?
Also the Doom benchmark is not only for comparison with other systems but to see how the game runs on certain hardware. So it would be wise to bench a screen size one uses in the actual game.

The difference between both screen sizes is not great and I don't want to split a hair, but it is really not understandable using the fullscreen mode.

I think I made clear whats my point. If you like, go ahead and bench in fullscreen mode that no one ever was using or is using even today.

no reason to be a cheney about it
also: i did not say your point was invalid but it seems that's how You understood me

I'm not a native english speaker and don't know what "cheney" means. I know a lot about "Dick Cheney" but don't see anything related to him here... 😀

Well, let's turn it around:
What is the positive reason to use a different screen size which makes comparability to all the results on old websites and newsgroups impossible?

i'm a naive engrish speakel either 😀
i was not looking for a positive reason, You may very well be right about the comparison stuff (if You trust results of other ppl), i'm just sayin it may not be so important anyway and if You later get a better FPS during gameplay than the benchmark says - that's usually a nice thing to happen;

Ages ago I put together this benchmark pack for a DOS Benchmarking project. I chose to use the most demanding setting in order to get good scaling on faster machines, fast Pentium II and III machines for example. Simple as that. I even mention this in the project brief

A classic and scales very nice as well. We will be using full screen to make the benchmark more demanding.

🤣

As for people using the benchmark pack, I guess it's because it's convenient. I want to get, especially new people, excited in retro gaming, so I put together some packs that make life easier. The DOS Starter Pack for example, or the pack for MS-DOS mode.

So people use the benchmark pack outside of the intended project, which I think is awesome:)

As for comparability with other results, the VGA Benchmark Database project: Phil's Ultimate VGA Benchmark Database Project

We have now collected over 700 entries, a really good effort from people from all around the world!

I always end up comparing to this database of results,

https://www.complang.tuwien.ac.at/misc/doombench.html
this are the settings they use

Get Doom version 1.9s and type in
doom -timedemo demo3
If you have already installed Doom, please restore the default settings (Graphic Detail: High; Screen Size: two steps below maximum (i.e., the status bar and one level of border)), as they strongly influence the results. Also, configure Doom for using keyboard and mouse (keyboard-only can give a speedup of 5% according to one report I received).