14 November 2013, 18:17 | #1 |
Registered User
Join Date: Nov 2009
Location: Legoland
Age: 45
Posts: 1,461
|
Dual channel could speed up WinUAE?
Dual channel is disabled on my computer, does this can hurt WinUAE performance greatly? How much, an estimative? To help you, its a Intel Dual Core E5200 2.60 GHz with 3 GB RAM DDR2.
Last edited by Leandro Jardim; 17 November 2013 at 21:38. |
15 November 2013, 09:30 | #2 |
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 49
Posts: 26,502
|
Answer is 100% yes, emulators need to move lots of data around, all the time, memory bandwidth is very important.
No idea how big or small is real world speed difference. |
15 November 2013, 10:22 | #3 |
Registered User
Join Date: Jul 2008
Location: Netherlands
Posts: 485
|
Hi Toni,
In case we would want to benchmark WinUAE performance, is there maybe some sort of benchmark option? I thought I read somewhere about "warp" mode, but I've no idea if that is meant as a benchmark option or how to activate it. Would be very nice to have a proper benchmark, gives us something more to play with when trying out different / new hardware |
15 November 2013, 10:51 | #4 |
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 49
Posts: 26,502
|
I am very anti-benchmark program. They can very easily show crazy results compared to real-world programs.
|
15 November 2013, 13:03 | #5 |
PSPUAE DEV
|
|
15 November 2013, 13:53 | #6 | |
Registered User
Join Date: Jul 2008
Location: Netherlands
Posts: 485
|
Quote:
For example my benchmark could be to run Hybris and Mental Hangover for a minute -unthrottled- in A500 quickstart most accurate. The result would be a % speed value, which tells you how many times faster the emulation is than a a real world machine running the same game or demo. The advantage being that it really tells something about the power of your hardware, and makes it comparable between different hardware setups. It would then for example also be easy to see (if someone would go through the effort) what the difference would be for single/dual/triple/quadruple, etc. memory bandwidths, or different cpu speeds, etc. I guess it could be a simple measure without the drawbacks of traditional benchmarking programs? You got my vote for a "-nothrottle" command line option at least |
|
15 November 2013, 14:10 | #7 |
Amiga 500 User
Join Date: Jun 2013
Location: EU
Posts: 1,501
|
Theoretically in WinUAE cycle-exact mode any benchmark program should show the same values as on a real Amiga, right?
Last edited by amilo3438; 15 November 2013 at 14:17. |
15 November 2013, 14:25 | #8 |
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 49
Posts: 26,502
|
It was about benchmark of your PC, how fast it can run emulation.
Just enable "warp" mode. |
15 November 2013, 14:30 | #9 | |
Amiga 500 User
Join Date: Jun 2013
Location: EU
Posts: 1,501
|
Quote:
EDIT: Or maybe to adjust the settings internally by itself with also showing some properly message. Such benchmark could test is a PC capable of running WinUAE at max load without dropping any frame of sound buffer or something else?! And continue to decrease some of option values, testing whole again until found the optimal value. EDIT1: And then generate "default" configuration file inside the same configuration folder. It would make life easier for beginners. EDIT2: Other possibility would be to run such benchmark after manually changing some option values to see if the PC is capable of running it at max load. p.s. If that all looks too much complicated, then the possible massage maybe would be f.e: "Buy the fastest = latest PC ... I am not going to support old PCs." Last edited by amilo3438; 15 November 2013 at 15:21. |
|
15 November 2013, 18:30 | #10 |
Registered User
Join Date: Jul 2008
Location: Netherlands
Posts: 485
|
|
15 November 2013, 18:38 | #11 |
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 49
Posts: 26,502
|
You missed my previous answer (and I think you still misunderstood the point of this test, it is not about capability to run emulation, it is about testing how much faster hardware X can run it compared to hardware Y or setting X vs Y)
Synthetic benchmarks are bad choice for this kind of testing. Users would attempt to optimize for case that never happens in real world and may actually slow down real world performance! end+pause or warp=true in config file. Frame limiter and sound gets disabled, emulation runs as fast as possible. |
15 November 2013, 18:58 | #12 | |
Amiga 500 User
Join Date: Jun 2013
Location: EU
Posts: 1,501
|
Quote:
(p.s. I know, it was always the question of which is better ... mean the PC. ) |
|
15 November 2013, 19:39 | #13 | |
Registered User
Join Date: Jan 2012
Location: USA
Posts: 372
|
Quote:
The biggest improvement in speed with WinUAE came after I installed a cheap PCI-e graphics card and turned off integrated graphics. I rarely see stuttering now. |
|
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
Thread Tools | |
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Create dual boot (AmigaOS/AmigaSYS) CF card with WinUAE | JackTheKnife | support.WinUAE | 14 | 29 November 2012 06:00 |
A600 dual kickstart, dual boot drive | TreacleWench | Hardware mods | 41 | 18 May 2012 12:02 |
WinUAE pad configuration (Dual Shock 2 Sony). | rudy1981_PL | support.WinUAE | 4 | 01 July 2008 12:24 |
WinUAE speed | frikilokooo | support.WinUAE | 5 | 30 September 2007 13:47 |
Running WinUAE on a dual-head system | anathema | support.WinUAE | 14 | 19 July 2003 15:09 |
|
|