About to jump into the pool

[Paladin]

 Wish me luck, as I'm about to dive into the GroovyMAME pool headfirst!
I built several MAME cabs years ago but then went the route of original hardware only.  After selling some of my original PCB's and cabs, I 'discovered' GroovyMAME' and how much closer to 100% emulation the current software and hardware is capable of.

I bought a converted War Final Assault cab with standard res WG monitor.  I'm just finishing a cap kit and flyback replacement.  I've also put together a PC from a couple Craigslist purchases as well as a new 120gb SSD.

Any comments on the PC build would be appreciated.  The goal was to stay low budget (I'm less than $200 into the PC), make a vertical monitor setup and be able to play as many of the Cave shmups as possible with minimum lag.
Specs:
1 x Case NZXT Tempest 410 Elite Gaming Case - Black
1 x Processor Intel® Core™ i7 970 Processor (6x 3.20GHz/12MB L3 Cache)
1 x Processor Cooling Liquid CPU Cooling System [SOCKET-1366]
1 x Memory 12 GB [4 GB X3] DDR3-1600 - G.Skill Ripjaws
1 x Video Card Radeon 6950 2gb
1 x Motherboard [SLI] ASUS P6X58-E PRO
1 x Power Supply 850 Watt Corsair CMPSU-850TXV2
1 x Optical Drive 24X Dual Format/Double Layer DVD±R/±RW + CD-R/RW Drive - Black
1 x Patriot 128gb SSD

I'm still reading up on the software side and am feeling a bit overwhelmed at the moment but I'm sure it'll all come together.

[schmerzkaufen]

I think you're gonna have to overclock that CPU a lot if your goal is to get the least possible lag.

Also 12GB RAM is way overkill, afaik 4GB is enough for a MAME machine.

Good luck! ^^

[krick]

I know the Cave games are very demanding.  I'm not sure if a 3.46 GHz (turbo) CPU will cut it but you might as well give it a shot and see.  If it isn't fast enough, you could always build a cheap system around the new i3-9100F CPU...  http://forum.arcadecontrols.com/index.php/topic,160485.0.html

Two bits of advice:

First, if your cabinet has a standard "JAMMA" harness, then pick up a J-PAC from Ultimarc...  https://www.ultimarc.com/jpac.html
It makes hooking everything up easier and it provides two useful benefits... 
1. It protects your monitor from out-of-range signals from the video card
2. It has a video amplifier

Second, if you're considering doing the Atom-15 video card firmware mod, don't bother until you've gotten everything working and stable.  And if you use a J-PAC, you really don't need Atom-15 anyway.  I usually just change my motherboard splash screen to a plain black image (using the motherboard manufacturer's provided tools) so I don't see anything away until Windows loads and CRT Emudriver kicks in.

[schmerzkaufen]

@krick: to be honest even with my CPU @4.2GHz the stable (100% no framerate drops) frame_delay level for overall cv1k Cave games I can afford is 5...
With some I can push to 6 and experience only a few drops during a run, but 7 is too disruptive and 8 out of the question.
Maybe my using Portaudio along increases the load, I'm not sure.
I plan to try 4.4~4.5GHz but I don't expect much difference, what about you? your CPU is more significantly faster.

I don't know how fast a CPU you need for a stable 7 or 8 with those games, not even mentioning 9 if that's even possible and might require a NASA supercomputer, but I doubt a CPU like the i3-9100F would produce results much different than my i5.

Still with frame_delay at 5 there's only 1/2 a frame or 8ms of delay left between you and the game, which isn't much at all if your hardware doesn't produce delay elsewhere (OS, controls, display) and already blazingly more responsive than standard MAME which adds like 3 frames if not worse.

I'm not sure but Paladin's CPU could maybe OC enough to achieve a similar experience, and who knows, maybe better.

[sharpfork]

I’ll be watching this thread as my requirements are similar. I have a rate cab and want to move to groovymame and be able to run SH3 cave games at the top end.

Are you thinking about windows or Linux?  Why?
I’m thinking Linux on my Tate can to keep it simple as it will only be mame.

I can start a different thread if preferred but figured out requirements are similar so I’ll post here.

I have:
i5 4670k
radeon hd 5450
Between 4gb-16gb of ram.

The big diff is that I have a 25” ntsc tv with component video in. I bought a retrotek vga to component transcoder to manage this interface.

[schmerzkaufen]

I’ll be watching this thread as my requirements are similar. I have a rate cab and want to move to groovymame and be able to run SH3 cave games at the top end.

I hope you're patient, because since I've been around here I've asked people with CPUs faster than mine what frame_delay / lag performance they achieve with cv1k/sh3 games, and as far as I remember I've never read a reply. *tumbleweeds* 

[Paladin]

Well, first off thanks for all the info so far!  Even if it's a bit disappointing on the speed front.

My cab was converted to a Tsunami MEGA arcade, basically a PC, audio amp, UVC to convert the PC VGA to standard res and an adapter board to take the controls and convert them to USB.  There is noJAMMA harness.

I read somewhere that the monitor may work without a video amp, the manual states it will work with 1.5v signals which still sounds a bit high but I'll try it and see once I get the video card set up.  Speaking of that, the card has dual bios.  I don't recall reading what's recommend in that case, flash both or leave one stock.

If I can get away with using a dvi to vga adapter, then a VGA cable to a breakout PCB wired to the monitor input, I'd only need to buy something for the controls as I have the video parts on hand.  The audio amp already takes mini stereo in and outputs to the stock shielded stereo speakers in the cab, so that's done already.

[schmerzkaufen]

it's a bit disappointing on the speed front.

Just to clarify if necessary; the Cave cv1k/sh3 games will very likely play at full speed on your system, it's the frame_delay lag reduction feature that might be limited (hence my suggestion to overclock)

[Recapnation]

I hope you're patient, because since I've been around here I've asked people with CPUs faster than mine what frame_delay / lag performance they achieve with cv1k/sh3 games, and as far as I remember I've never read a reply. *tumbleweeds* 

Probably has to do with how far from optimal/finished is that system's emulation in MAME. Worrying about the additional < 17 ms of lag the frame delay feature would save with high-end CPUs when the games just don't slow down when they should --an essential aspect in games like these-- doesn't seem to follow exactly the most compelling rationale. (Doesn't MAME show the non-[properly] working flag for these games?)

[schmerzkaufen]

Probably has to do with how far from optimal/finished is that system's emulation in MAME. Worrying about the additional < 17 ms of lag the frame delay feature would save with high-end CPUs when the games just don't slow down when they should --an essential aspect in games like these-- doesn't seem to follow exactly the most compelling rationale. (Doesn't MAME show the non-[properly] working flag for these games?)

Sure, however Groovy is the only modern build that saves both the CPU% and Blitter sliders that can help reproduce those slowdowns to a certain degree (not accurately like the actual boards but in a flavor comparable to 360/PC ports)

Though people seem to place the question of input delay first as something to take care of before researching the best cpu+blitter values (which takes a significant amount of research time but that's not the topic yet)
And in order to answer that question it is understandable that they'd want to know a bit about what lag/frame_delay performance they should expect with X or Y magnitude of processing power.
Because we're not talking about trying a handful of cheap second hand old GPUs here (which don't play a big role there anyway), but about investing in a powerful - and therefore expensive - CPU.

I've been wondering myself what CPU could provide a stable frame_delay 8 (assuming even using HLSL) and I still have no idea, as for instance I couldn't push my current CPU beyond 4.3GHz and it didn't make a noticeable difference.
It's natural that I would ask people with 4.8, or even 5GHz monsters how it fares for them.


On an additional note, there are people doing funky things like this: http://electricunderground.io/shmup-input-lag-database/ (Groovy measurements down the list)
which makes me frown because although I haven't made live delay measurements myself, I certainly don't feel that I'm playing Futari with 5 frames of lag in Groovy.
besides Futari the guy might have mis-configured everything anyway, seeing how bizarre it is to turn sync off with Groovy. welp.

[Recapnation]

Sure, however Groovy is the only modern build that saves both the CPU% and Blitter sliders that can help reproduce those slowdowns to a certain degree (not accurately like the actual boards but in a flavor comparable to 360/PC ports)

That's not saying much, I'm afraid. The X360 and WIN ports vary from okay-ish to intolerable in this regard. Surely with the slider options in MAME you can mimick the original behaviour much better than most of the Cave ports do. But it seems nobody is concerned enough to make the effort and share the results after all this time, so bothering about 1 more frame of lag (and investing in a high-end CPU for this particular matter) is kind of pointless, if you ask me.

On an additional note, there are people doing funky things like this: http://electricunderground.io/shmup-input-lag-database/ (Groovy measurements down the list)
which makes me frown because although I haven't made live delay measurements myself, I certainly don't feel that I'm playing Futari with 5 frames of lag in Groovy.
besides Futari the guy might have mis-configured everything anyway, seeing how bizarre it is to turn sync off with Groovy. welp.

Looks like a commendable effort but some things are indeed questionable. If he's getting "100 % consistency" with frame delay at 6 he's either, not using GM properly or not taking enough samples, unless I'm missing something. Not posting his PC, Windows and GM configurations also makes his effort futile in these cases. And we don't know how "good" the controller he's using on Windows for this is, either (does he?).

I don't find the figures anything uncommon, anyhow. If those Sega Saturn and PS2 games delay 3 frames for him (which I always found to be perfectly responsive), the GM figures are basically optimal -- I'd hardly believe that the CV-1000 games' native lag is below or the same as PS2 Dai-Ou-Jou or SS Batsugun, for instance. That's the other big issue of that chart -- it lacks tests of the CV-1000 PCB counterparts. (And, for the only PCB he measured, he didn't bother to make the comparison with GM. (!))

[schmerzkaufen]

Surely with the slider options in MAME you can mimick the original behaviour much better than most of the Cave ports do. But it seems nobody is concerned enough to make the effort and share the results after all this time

I've been interested in that for quite a while but it's too much for one person, it would have to be a common effort/contribution. Quite a bit were interested a few years ago but two-three changes in MAME kinda pushed the issue to later (a couple driver performance optimizations, and the removal of saving CPU slider values).
Yet after the optimizations were done, and now the sliders saving is back in GM, it seems too much time has passed. I still frequently see people interested in playing the cv1k in emulation as well as possible, but either they're bewildered by Groovy's rumored complexity (don't ask me why, for that purpose there isn't much to do and configure, yet more than a couple settings seem to scare people), or they're not aware about the sliders, or they just don't believe it.

Wasted opportunity IMO, not showing off what Groovy+frame_delay can achieve delay-wise included.
AFAIK some compatibility issues with baseline MAME in RetroArch might be solved soon, which might mean run-ahead finally working for them, if that's confirmed then you can bet no one who was interested in Groovy for playing those cv1k games will even consider trying anymore, they'll just go to RA and push run-ahead mindlessly.

Looks like a commendable effort but some things are indeed questionable. If he's getting "100 % consistency" with frame delay at 6 he's either, not using GM properly or not taking enough samples, unless I'm missing something. Not posting his PC, Windows and GM configurations also makes his effort futile in these cases. And we don't know how "good" the controller he's using on Windows for this is, either (does he?).

Not sure, the stick is a VX-SA though, allegedly the fastest known stick  according to that website: http://www.teyah.net/sticklag/results.html, but in windows (and which windows) I dunno.

I don't find the figures anything uncommon, anyhow. If those Sega Saturn and PS2 games delay 3 frames for him (which I always found to be perfectly responsive), the GM figures are basically optimal -- I'd hardly believe that the CV-1000 games' native lag is below or the same as PS2 Dai-Ou-Jou or SS Batsugun, for instance. That's the other big issue of that chart -- it lacks tests of the CV-1000 PCB counterparts. (And, for the only PCB he measured, he didn't bother to make the comparison with GM. (!))

Not sure for all that, I was taking Futari as reference since according to M2 the pcb lags by the equivalent of 2 frames, and if you remember the 360 port offers to adjust your delay to that. That's what he measured as well but with VSYNC OFF...why? it's supposed to stay vsynced and the delay still be that, so that's how it should have measured.
There's belief probably inherited from the PC gaming world that vsync always adds significant lag no matter what, but when you try to explain people that something like what Groovy does allows you to keep the sync, yet keep the cost under 1 frame, down to even mere 2~3ms under conditions, they simply don't believe or don't understand what it means (like some of those RA/ShmupArch users who think even the game's default lag is input delay to eliminate period *shrug*)

In any case, we can measure the driver's default delay in Groovy (does the driver lags by 2 or 3 frames by default or is it different per-game) that's something I'll probably do when I have the time to make sure it's done properly, and then deduce what can be achieved with frame_delay.
But that still doesn't inform about the upper limits nor what one can hope to achieve with what CPU...
Really it doesn't have to be 9, but a stable 8 or even 7 puts you in the decently achievable 'under half-a-frame-delay' zone which is like having no input lag at all assuming your hardware/OS are clean of most additional delay.

[digitron]

Really it doesn't have to be 9, but a stable 8 or even 7 puts you in the decently achievable 'under half-a-frame-delay' zone which is like having no input lag at all assuming your hardware/OS are clean of most additional delay.

Really appreciate your post Schmerzkaufen, as I finally got my Vertical GM setup up and running, it feels pretty good out of the box yet I wonder if there is any room for improvement? Are you referring to the Windows Experience Index in Windows 7 64 bit? (I'm at 7.3 = i5-2400 @ 3.1Ghz, SSD, ATI HD 5750, 8GB RAM) Are there any recommended setting changes for Cave/Shmups for GM or should I just leave it alone?

[schmerzkaufen]

Nope I'm referring to 'frame delay', it's a feature used to mitigate input lag further than what Groovy produces by default.
You'll find it in game; press TAB, go to the sliders menu, and adjust it while playing until you find a stable value/behaviour (press F11 to check the framerate always remains at 100%), there's 9 levels, the higher the more responsive, but also the more CPU intensive.

EDIT: and if you want to give simulating slowdowns a try, open the mame.ini and turn the 'cheat' line on (write 1 instead or 0 then save). This will make new sliders appear, the Blitter and CPU are the ones you'll need.
Back when I was trying I remember getting decent results in some games like Pink Sweets reducing the CPU down to +/- 40%, and the Blitter to 63% (some say 59%), but those values can be completely wrong in most games, it's mostly unexplored tweaking territory.

[digitron]

Nope I'm referring to 'frame delay', it's a feature used to mitigate input lag further than what Groovy produces by default.
You'll find it in game; press TAB, go to the sliders menu, and adjust it while playing until you find a stable value/behaviour (press F11 to check the framerate always remains at 100%), there's 9 levels, the higher the more responsive, but also the more CPU intensive.

EDIT: and if you want to give simulating slowdowns a try, open the mame.ini and turn the 'cheat' line on (write 1 instead or 0 then save). This will make new sliders appear, the Blitter and CPU are the ones you'll need.
Back when I was trying I remember getting decent results in some games like Pink Sweets reducing the CPU down to +/- 40%, and the Blitter to 63% (some say 59%), but those values can be completely wrong in most games, it's mostly unexplored tweaking territory.

Wow, lots of cool knobs here, that you for the expert reply.

Just tried DoDonPachi, I can only increment the 'frame delay' by 1 while maintaining 100% frame-rate. When I go to '2', it begins to fluctuate in-between 100% and 94%.

Oh I see it now, CPU was definitely getting maxed out when I turned it all the way up, lol.

[schmerzkaufen]

^ Could be your settings, could be your CPU can't handle even a little of the frame delay feature (what's your cpu? edit: sorry i've seen it, yeah it might be too weak, dunno for sure but look at my sig: I can only afford frame delay 5 for the cv1k, but easily 7~8 for the older cave titles like DDP)

And for the other sliders (cpu% and blitter) note they're only for the 'CV1000' a.k.a 'cv1k' a.k.a 'SH3' hardware-based Cave games, meaning Mushihimesame up to Akai Katana (DOJ not being emulated in MAME yet)

[digitron]

^ Could be your settings, could be your CPU can't handle even a little of the frame delay feature (what's your cpu? edit: sorry i've seen it, yeah it might be too weak, dunno for sure but look at my sig: I can only afford frame delay 5 for the cv1k, but easily 7~8 for the older cave titles like DDP)

And for the other sliders (cpu% and blitter) note they're only for the 'CV1000' a.k.a 'cv1k' a.k.a 'SH3' hardware-based Cave games, meaning Mushihimesame up to Akai Katana (DOJ not being emulated in MAME yet)

You only need an i5 for GroovyMame they said, lol! Fun stuff, you're teaching me a lot here.

I'm running GM v. 208 if that makes a difference?

Here is after I exited GM with 1 frame delay set in DoDonPachi. It's not using my entire CPU, only a single core I take it?

[schmerzkaufen]

Your i5 can turbo to 3.4GHz, that seems to be enough for running all Cave games at 100%, but using frame_delay on top requires a CPU as fast as possible, yeah.

GM v0.208 ... mmh I don't know, Calamity and Doozer were talking about shoving a sync library in or something that would make the feature more accurate but a bit more demanding. I don't know the status.

EDIT: for the CPU usage I don't know, last time I've checked when I play all 4 are used more or less, no flat line.

[digitron]

Did a few tests and it seems most new Cave games drop frames even with 1 frame delay enabled, I mean they all feel good play-wise so I dunno. Maybe I'll just leave it for now.

The more I learn about GM the more I really like it, finding up to date documentation has been my struggle and I feel guilty bringing it up as I know Calamity is a one-man band these days. It's a super project and I'm just getting my feet wet!

[schmerzkaufen]

Wait a minute, I remember now, v0.208 had issues with stability, but that was baseline MAME's fault.

Try again with 0.210

[digitron]

Wait a minute, I remember now, v0.208 had issues with stability, but that was baseline MAME's fault.

Try again with 0.210

Ok, Ugh, I hate downloading the huge romsets, lol. Thanks

[schmerzkaufen]

If you have a 0.208 romset it is likely that 99% of you roms will work with 0.210 too. Just grab the latest Groovy and you're good.

You can alway find the few missing roms later.

[digitron]

If you have a 0.208 romset it is likely that 99% of you roms will work with 0.210 too. Just grab the latest Groovy and you're good.

You can alway find the few missing roms later.

That was it! Thank you.

As a test I have frame delay set to 8 (DoDonPachi) and no frame drops, I'm not sure if if this is optimal. I re-installed the modes with VMMaker and let it do all the vsync handling, etc. CPU sits at around 20% so definitely a change from 208 to 210.

Mushihimesame = 7 frame delay, zero frame drops, CPU ~40% utilization

So is this a good methodology for getting the least amount of lag for Shmups in general? Cave/vertical shmups are the only genre I care about in GM so any other titles I don't care about. Hit F11 and increase frame delay until I get zero frame drops?

Thanks again.

[schmerzkaufen]

Good! What about cv1k games ? Mushihimesama, Pink Sweets etc, how far can you push frame delay with those ?

edit: cross-edits lol. Yes that's a good method, you can adjust while playing and it's saved per-game every time so no worries.

7~8 is very good, the absolute best being 9 but it's rarely manageable.

[krick]

Ok, Ugh, I hate downloading the huge romsets, lol. Thanks

You can just start a new torrent for the 0.210 ROM set, then completely exit your torrent client, copy all the 0.208 ROM files into the 0.210 folder, then start your client up and force a recheck.

I'm in the process of doing this right now going from 0.203 to 0.210 and 90%+ of the ROM files are the same.

Just make sure that you keep the same format...  merged to merged, split to split, etc...  for best results.

[digitron]

Good! What about cv1k games ? Mushihimesama, Pink Sweets etc, how far can you push frame delay with those ?

edit: cross-edits lol. Yes that's a good method, you can adjust while playing and it's saved per-game every time so no worries.

7~8 is very good, the absolute best being 9 but it's rarely manageable.

Awesome!

So far..

Muchi Muchi Pork - 7
Pink Sweets - 7
Mushihimesama - 7
Espgaluda II - 7
Mushihimesama Futari / 1.5 / Black Label - 9

Ok, Ugh, I hate downloading the huge romsets, lol. Thanks

You can just start a new torrent for the 0.210 ROM set, then completely exit your torrent client, copy all the 0.208 ROM files into the 0.210 folder, then start your client up and force a recheck.

I'm in the process of doing this right now going from 0.203 to 0.210 and 90%+ of the ROM files are the same.

Just make sure that you keep the same format...  merged to merged, split to split, etc...  for best results.

Fantastic, great to know!

[haynor666]

Wish me luck, as I'm about to dive into the GroovyMAME pool headfirst!
I built several MAME cabs years ago but then went the route of original hardware only.  After selling some of my original PCB's and cabs, I 'discovered' GroovyMAME' and how much closer to 100% emulation the current software and hardware is capable of.

I bought a converted War Final Assault cab with standard res WG monitor.  I'm just finishing a cap kit and flyback replacement.  I've also put together a PC from a couple Craigslist purchases as well as a new 120gb SSD.

Any comments on the PC build would be appreciated.  The goal was to stay low budget (I'm less than $200 into the PC), make a vertical monitor setup and be able to play as many of the Cave shmups as possible with minimum lag.
Specs:
1 x Case NZXT Tempest 410 Elite Gaming Case - Black
1 x Processor Intel® Core™ i7 970 Processor (6x 3.20GHz/12MB L3 Cache)
1 x Processor Cooling Liquid CPU Cooling System [SOCKET-1366]
1 x Memory 12 GB [4 GB X3] DDR3-1600 - G.Skill Ripjaws
1 x Video Card  2gb
1 x Motherboard [SLI] ASUS P6X58-E PRO
1 x Power Supply 850 Watt Corsair CMPSU-850TXV2
1 x Optical Drive 24X Dual Format/Double Layer DVD±R/±RW + CD-R/RW Drive - Black
1 x Patriot 128gb SSD

I'm still reading up on the software side and am feeling a bit overwhelmed at the moment but I'm sure it'll all come together.

Certainly I will fall into disfavour some people here but it was many time repeated - use processor with most powerfull single core instead with many cores. i7 970 is in my opinion for 90% in mame not the best choice, it's old and very power hungry.
Another problem is 12 Gb of RAM. Total waste for mame or many other emulators. 4 Gb is enough. I have system with 2 Gb and still everything is fine (or at least it was with groovymame around version 203, right now I'm testing version 210 in my PC).
PSU = do You need so much power even with such card and processor?
Graphics card - an 200W monster :/
Optical drive is naturally not needed.
You ssd is fine but size depends how many and what games do You need. I personally some time ago bought 256 Gb SSD because 128 Gb was not enough for selected mame games (including some laser games), Taito Type X/X2 games, naomi games (emulated in nullDC), NES anf GBA games. For lazy people that copy paste whole set 600 Gb would not be enough

Here where I live (in POland) I've found Dell 3020 with i3-4130 that should be in my opinion much better choice. It cost about 140 USD. All You need is some low power Radeon for groovymame which will cost around 30 USD. krick has very good hardware (look at his signature) but I'm affraid it still might be expensive.

You can just start a new torrent for the 0.210 ROM set, then completely exit your torrent client, copy all the 0.208 ROM files into the 0.210 folder, then start your client up and force a recheck.

I'm in the process of doing this right now going from 0.203 to 0.210 and 90%+ of the ROM files are the same.

Just make sure that you keep the same format...  merged to merged, split to split, etc...  for best results.

Use of clrmamepro would be even better choice though a bit more complicated.

[schmerzkaufen]

Certainly I will fall into disfavour some people here but it was many time repeated - use processor with most powerfull single core instead with many cores.

Lol we'll go in circles around that point forever, to be clear nobody denies that single core performance/Ghz is the most important aspect, but having more cores is definitely an advantage for the games that do use them, which are more than you think. 

[haynor666]

Majority of games in mame still relies on single core. For now known games to work better on multicore machines list is short.

In this case however if Paladin need cv1k games to work as best as possible 4 core processor might be better choice.

Here are my results on G3258 that have score in passmark 2500 points using official mame 205 x64 and -bench 299:

futaribl   686,62
futari15   716,77
mmpork   841,83
espgal2   910,13
dfkbl   1002,45
pinkswts   1085,94
dsmbl   1104,23
ibara   1109,41
ibarablk   1138,31
mushisam   1173,40
deathsml   1195,52
ddpdfk   1206,35
akatana   1227,62
mushitam   1235,26

[Sledge]

Regarding building a PC, i've found the Dell 9010/9020 to be very good and cheap for mame, just add a suitable GPU and you're good to go..
Mine are i5-3570 with SSD etc and only cost me about $200 each

[Paladin]

 I bought a used PC with no HD and a bad graphics card, so the components were all included and not something I had a choice about.  The used 6950 cost me $20 so I figured what the heck.

I thought perhaps the DDR3 RAM configuration might squeeze a bit more performance out of the system, and agree that 12gb is overkill.  I'm happy with the power supply though, as it should last longer if it's not driven to max specs.

It's all a fun experiment anyway and if this system doesn't get me to where I'm satisfied, it's still a very nice rig to put to other uses.

Now, what's the state of input devices?  Is Ultimarc still on top, or are there better options for PC to arcade controller now?

[haynor666]

Regarding building a PC, i've found the Dell 9010/9020 to be very good and cheap for mame, just add a suitable GPU and you're good to go..
Mine are i5-3570 with SSD etc and only cost me about $200 each

Yeah, I'm writting this message from Dell 9020 at my work place. It has i5-4690. Really stable and was already tested with mame

[schmerzkaufen]

I like that some of those refurbished Optiplex come with a 3+GHZ quad core and W7, sometimes for +/- 200 indeed.

So far..

Muchi Muchi Pork - 7
Pink Sweets - 7
Mushihimesama - 7
Espgaluda II - 7

You know what's funny? I'm here all about "get moar Hz not just for emulation but also taking frame_delay into account", and then you see results like yours that tell the CPU speed might not be the most determining factor after all.

Some time ago I was curious about what a more powerful GPU does to Groovy since it was said there was an advantage, so I've acquired a handful from entry AMD R7's up to former mid-ranges like the R9 380X and RX 470 (half-intendedly eBay being the mess it is lol), and what I've found is that they barely make a difference, they matter mainly to HD displays users like me, and only in occurences where you need a bit more juice for eliminating tearing or when you use HLSL. Verdict? with the R9~RX on average a 'gain' of 1 level of frame_delay...

So I've been thinking we don't really know what makes frame_delay perform better or worse, it was said that it is mainly the CPU but that's probably only a portion of the story.
That's why I'm disappointed that more people don't share such performance results with their own - sometimes quite powerful - hardwares, sure it's easy to tell people to get a decently powerful CPU like a bit over 3GHz and a low end AMD with analogue out since that will work for most games and displays. But for those who venture into heavier games emulation, and seek optimal lag reduction, or use high-resolution displays, it's mostly mist and shadows (and money spent sometimes in vain...)

[digitron]

I like that some of those refurbished Optiplex come with a 3+GHZ quad core and W7, sometimes for +/- 200 indeed.

So far..

Muchi Muchi Pork - 7
Pink Sweets - 7
Mushihimesama - 7
Espgaluda II - 7

You know what's funny? I'm here all about "get moar Hz not just for emulation but also taking frame_delay into account", and then you see results like yours that tell the CPU speed might not be the most determining factor after all.

Some time ago I was curious about what a more powerful GPU does to Groovy since it was said there was an advantage, so I've acquired a handful from entry AMD R7's up to former mid-ranges like the R9 380X and RX 470 (half-intendedly eBay being the mess it is lol), and what I've found is that they barely make a difference, they matter mainly to HD displays users like me, and only in occurences where you need a bit more juice for eliminating tearing or when you use HLSL. Verdict? with the R9~RX on average a 'gain' of 1 level of frame_delay...

So I've been thinking we don't really know what makes frame_delay perform better or worse, it was said that it is mainly the CPU but that's probably only a portion of the story.
That's why I'm disappointed that more people don't share such performance results with their own - sometimes quite powerful - hardwares, sure it's easy to tell people to get a decently powerful CPU like a bit over 3GHz and a low end AMD with analogue out since that will work for most games and displays. But for those who venture into heavier games emulation, and seek optimal lag reduction, or use high-resolution displays, it's mostly mist and shadows (and money spent sometimes in vain...)

I ran through a few more games last night and I had Mushihimesama Futari BL stable at 9 frame delay. I'm running a 5750 card, i5 3.1ghz, crapper free PC.

Noticed that even when I get frame drops, my CPU utilization rarely goes above 50%, it does increase but I'll get frame drops before my CPU maxes out.

[schmerzkaufen]

Go figure, in my case I understand that using a high resolution display I'm supposed to need more power, but I don't understand why I don't get any better results at all in OC vs. the base 3.5~3.9 clock even when I push 400 more Mhz (4.3GHz) not even just 1 additional frame_delay step is allowed. I might try 4.4 but I think that's the limit for my PC as it is.
(my CPU doesn't get much usage either, but I think it's only indication of the emulation and the frame_delay toll doesn't translate in the graphs here)

Maybe that means frame_delay scales extremely badly, that you easily benefit when you're using a CRT, but you're very quickly bottlenecked with a HD, Full-HD, or WQHD, and I don't even want to imagine what happens with a 4K. If that's the case then I can't guess what kind of race horse I need for my future config if aiming for at least 3 more comfortable steps.

[krick]

I was in the process of updating my cabinet to the latest GroovyMAME 0.210 over the weekend with intentions of doing some benchmarking and testing with frame_delay on my hardware and then... my trusty old Hantarex Polo 25 arcade monitor died.   

Hopefully, I'll be able to get it working again.

[digitron]

I'm running DirectX v. 11, do you guys specifically uninstall that and install the DirectX v. 9 for Groovy? Does it make a difference? Things seem to be running ok on 11?

[Calamity]

I was in the process of updating my cabinet to the latest GroovyMAME 0.210 over the weekend with intentions of doing some benchmarking and testing with frame_delay on my hardware and then... my trusty old Hantarex Polo 25 arcade monitor died.   

Hopefully, I'll be able to get it working again.

Mate, so sad to hear that. Sure you'll get it back in order.

[krick]

I'm running DirectX v. 11, do you guys specifically uninstall that and install the DirectX v. 9 for Groovy? Does it make a difference? Things seem to be running ok on 11?

You can't uninstall v11.  But you should definitely install the full DirectX 9 redist.

Here's the full offline installer...

https://www.microsoft.com/en-us/download/details.aspx?id=8109

It should have downloaded this file...

directx_Jun2010_redist.exe

When you run the file, it will ask you where to extract it.  Extract it somewhere with a simple path like C:\dx9redist

Once it extracts, go into that folder and run this file...

DXSETUP.exe

[digitron]

I'm running DirectX v. 11, do you guys specifically uninstall that and install the DirectX v. 9 for Groovy? Does it make a difference? Things seem to be running ok on 11?

You can't uninstall v11.  But you should definitely install the full DirectX 9 redist.

Here's the full offline installer...

https://www.microsoft.com/en-us/download/details.aspx?id=8109

It should have downloaded this file...

directx_Jun2010_redist.exe

When you run the file, it will ask you where to extract it.  Extract it somewhere with a simple path like C:\dx9redist

Once it extracts, go into that folder and run this file...

DXSETUP.exe

Thanks!