PC-Task vs. Bridgeboard

[YouKnowWho]

Back in the day, this was much more clear. But today, with many running CPU upgrades, in many cases 68060, what exactly is the performance of PC Task?
Say you have a 50Mhz 68060 in a 1200 or 25Mhz 68040 in a 4000. How would that PC-Task setup perform in PC speed tests? What would it compare to PC CPU speed wise? Surely it would beat a A2286AT or maybe even A2386SX, right? And it doesn't require a big-box Amiga and board? In case of 4000D precious slot(s) taken up too. I have seen vidoes on YouTube of PC DOS games on 68060 Amigas, but not test results showing what performance level one can expect with PC Task emulated PC. How fast is that emulated 486? Anyone have any links or test results or info? Thank you.

EDIT:
Look at these 1200/030 50Mhz tests - quite decent already. Would 060 50Mhz deliver a 486 of 25Mhz?
FYI: https://retrocomputing.stackexchange.com/questions/17074/pc-task-performance

 

[comraider]

Look at these 1200/030 50Mhz tests - quite decent already. Would 060 50Mhz deliver a 486 of 25Mhz?

That might be pushing it.. Benchmarks are tricky here of course, but this old video claims to show an 060 90MHz coming in around there, although it runs Windows 95 very sluggishly:

 

[Xanxi]

I copy paste one of my old post in another thread from a couple years ago.

"Finally, for reference here are some benchmark results of Norton Sysinfo:


* the GG386 AMD 386sx @25MHz



* The Commodore A2386 with Intel 386sx @ 25MHz - faster than GG386



* The Commodre A2386 with Cyrix 486SRX2@50 MHz - almost a 486 DX33




* PCTask with 68060@50 MHz pretending to be a 486sx cpu, roughly equivalent to a 386sx@25 (but the video emulation is slow even with a lot of RAM and Z3 RTG board so it is nowhere as smooth as any 386 bridgboard)

"

 

[YouKnowWho]

@Xanxi - YES! THANK YOU! That title being specific to GG board perhaps resulted in it not coming up in my search. Thank you. SO GOOD!

May I ask if this was a fair comparison? I mean, all the bridgeboards are CGA 2 or 4 colour stuff and look interlace/different res, but the PC-Task scren looks crisp, as if a higher quality video setting was used with PC-Task test, not basic CGA for example. OR was it? Could you confirm reason for the better looking PC-Task screen vs. bridgeboards? Any difference if you run CGA (bridgeboard equivalent) video option on PC-Task on the native chipset (I'm guessing 4000) and again on the Z3 RTG board? Is the Amiga workbench screen high res and 24bit? ...as in any resource demands loading up the system? I mean...if better video emulation options are available on PC-Task, who wouldn't be tempted to use them?

Finally, can you please validate first test and second test on PC-Task if you have a chance in the future. Is there an improvement once things get into memory and are run the test a second time? ...just wondering.

I'm in the "putting together stuff" stage right now of a reasonable setup for this and I'd like to tweak for PC-Task, if at all possible. Obviously lots of RAM, but if RTG makes little difference, I'll live with fewer colours. I'm also trying to understand the ideal PC-Task set up choices. Found that "The great PC-Task thread" as well that I'm reading now. Thank you for your details @Xanxi!

 

[Xanxi]

I use the bridgeboards (the upgraded Commodore A2386 being my favorite and stay inside my main A4000) only with dedicated VGA card, sound cards and so forth. Those tests were performed with the STB Nitro 2 MB ISA which is one of the best card around. Pictures were taken for the bridgeboards from the VGA input of my monitor, with a scanline generator connected in between the VGA card and the VGA input, whereas PCTask screenshot is taken from the HDMI input of the same monitor out of the ZZ9000 with no scanlines.
Norton Sysinfo uses only a few colors anyway and run in 80 columns mode. For each PCTask resolution, i use a matching or closest resolution from the ZZ9000 P96 in 8 bit, which is the minimum color depth i can set and is fine for PCTask from 2 to 256 colors. There is not better to expect from AGA rather than Z3 RTG. In my experience of video emulation for PCTask or Shapeshifter in 256 colors or less, AGA outperforms any Z2 RTG card, but is of course beaten by Z3 bandwitch.

Norton Sysinfo always gives the same result with the BB once set up for good. PCTask can fluctuate a bit its score around 0,1 or 0,2.
In real use, just test Dune.
The game runs perfect with a BB A2286, is more than perfect with 386sx bridgeboards. PCTask on 68060 is killed by Dune as soon as there is the cycling color effect in the intro and crawls like hell. PCTask is really a clean program, rarely crashes on a 68060 but can't do more than productivity stuff under MSDOS or very old games with beeper (Monkey 1 EGA not bad). On Vampire V2 or Pistorm it tends to become instable on my setup, but i have seen Youtube vids of someone running them with Vampire + and a GoldenGate II + Adlib card combo.

 

[Magic]

@ Mr Xanxi​

This is great information, thank you for taking the time to post it!

 

[YouKnowWho]

@Xanxi, indeed - details are very much appreciated. Well, it seems for starters that any BB below A2386SX is going to be rivalled by PC-Task. Of course, no BBs for wedges, so PC-Task is a lovely thing to have vs. a A2286AT. Oh...have you tried PCx? I saw a video that claims it is faster/better than PC-Task. (Where can one purchase PCx anyway?)



The thing about BBs with a VGA card is for me is that at that point, you're switching and don't have an option of video in the WB window anymore, a plus for PC-Task. By the way...any faster when PC-Task is in a window vs. screen or none at all? Forgive for all the questions, answers are much appreciated. I'm a measure many times kind of a guy, then cut.

 

[Flash951]

I've used PCTask a lot, also BB. I currently have A2088XT (modded to 9.5 MHz) setup in A2000, A2286 (10 MHz modded in another A2000, and A2386SX (60 MHz 486, 12 mhz isa) in my A4000 060@100. BB beats PCTask any time in real life. But PCTask is great for productivity on a fast Amiga. If you don't want to add VGA, maybe IDE card and stuff to BB, you are stuck with CGA, which are good enough for productivity in DOS (ANSI). If you got a RTG card you can use Viaduct and have CGA in a window etc. PCx is unreliable, like a beta version.

 

[YouKnowWho]

@Flash951. Yes, it seems no matter the solution to run PC on Amiga, there are compromises. Interesting feedback on PCx, I won't waste time on it then. PC-Task being solid appears to be the concencus. Any setting suggestions for optomization of PC-Task form your experience? Do you just use AGA or RTG? Window or Screen? Obviously partition not file. Suggestions much appreciated.

I see we have one BIG thing in common here about the PC inside the Amiga - desire for performance. I see beefed-up BBs here being a common thread. I see everyone push their BBs and add video and audio cards. Thinking back, I didn't even give PCs attention until 486DX66 because my Amiga 2000/040 was so awesome. It took Pentium II MMX to get me to consider a PC. I think PC-Task will do the trick for "up to 486" needs, productivity mainly indeed.

 

[Flash951]

@Flash951. Yes, it seems no matter the solution to run PC on Amiga, there are compromises. Interesting feedback on PCx, I won't waste time on it then. PC-Task being solid appears to be the concencus. Any setting suggestions for optomization of PC-Task form your experience? Do you just use AGA or RTG? Window or Screen? Obviously partition not file. Suggestions much appreciated.

I see we have one BIG thing in common here about the PC inside the Amiga - desire for performance. I see beefed-up BBs here being a common thread. I see everyone push their BBs and add video and audio cards. Thinking back, I didn't even give PCs attention until 486DX66 because my Amiga 2000/040 was so awesome. It took Pentium II MMX to get me to consider a PC. I think PC-Task will do the trick for "up to 486" needs, productivity mainly indeed.

PCTask will NOT give you anywhere near 486, only in Sysinfo and in your imagination, as we have tried to explaine.

I use vga and ide io card only in my a4000/60@100mhz. For the others I stick to cga on Amiga screen with BB.

PCTask: hardfile works very well, just use pfs or a fast file system, but not fastfilesystem - it's not.

Window is slow, use screen. Use rtg scrren for 256 colors, for less, use Amiga chipset, read manual about why.

 

[YouKnowWho]

PCTask will NOT give you anywhere near 486, only in Sysinfo and in your imagination, as we have tried to explaine.

I use vga and ide io card only in my a4000/60@100mhz. For the others I stick to cga on Amiga screen with BB.

PCTask: hardfile works very well, just use pfs or a fast file system, but not fastfilesystem - it's not.

Window is slow, use screen. Use rtg scrren for 256 colors, for less, use Amiga chipset, read manual about why.

Will do. Of course, I fully understand from the spec tests and notes above that it will not be like a real 486SX, and certainly not for anything graphics intensive. But it reads like PC-Task can be nearly stock A2386SX-ish. Certainly not for any fancy games or graphics, even if it can do more colours than BB in a window or screen as a convenience. Appreciate the PC-Task setup advice and saved time not wasted on PCx as well.

EDIT: I'd like to see a competition to get the best Norton Sys-Info score for PC-Task using a real 68K CPU.

FYI - A2286AT screen grab of CPU speed test from a YouTube Video of a 2000 with A2286AT with VGA card installed. Video is low res.

 

[Flash951]

Will do. Of course, I fully understand from the spec tests and notes above that it will not be like a real 486SX, and certainly not for anything graphics intensive. But it reads like PC-Task can be nearly stock A2386SX-ish. Certainly not for any fancy games or graphics, even if it can do more colours than BB in a window or screen as a convenience. Appreciate the PC-Task setup advice and saved time not wasted on PCx as well.

EDIT: I'd like to see a competition to get the best Norton Sys-Info score for PC-Task using a real 68K CPU.

FYI - A2286AT screen grab of CPU speed test from a YouTube Video of a 2000 with A2286AT with VGA card installed. Video is low res.

View attachment 2547248

Its my video you found on YouTube btw, so subscribe and like. But what you seem to not comprehend is that Sysinfo CPU speed in PCTask is just bs, it shows much faster results than actual emulation performance. Probably its a short code with same instructions repeated over and over and PCTask compile and cache this small code in 68k code and it gives a very bogus result.

 

[YouKnowWho]

@Flash951, I understand that benchmark software can be gamed. Still, PC-Task seems quite usable as a gap measure and to satisfy that itch of running something from PC on the Amiga. @Xanxi 68060 50Mhz test shows 4x A2286 speed on the CPU Speed test, is PC-Task embellishing the test results that much? It's more fun than any critical function at this point as we don't need a working retro PC for our grades or work today. I'm still surprised that I can't find a result of someone with an 060 90/100/105Mhz Amiga out there running a CPU Speed test on their PC-Task. Could you perhaps do a grab a PC-Task CPU Test on an 060/100Mhz or do a new video with that? I think it would get some views. Cool that I would find a useful video and it would be yours @Flash951 !

 

[Flash951]

@Flash951, I understand that benchmark software can be gamed. Still, PC-Task seems quite usable as a gap measure and to satisfy that itch of running something from PC on the Amiga. @Xanxi 68060 50Mhz test shows 4x A2286 speed on the CPU Speed test, is PC-Task embellishing the test results that much? It's more fun than any critical function at this point as we don't need a working retro PC for our grades or work today. I'm still surprised that I can't find a result of someone with an 060 90/100/105Mhz Amiga out there running a CPU Speed test on their PC-Task. Could you perhaps do a grab a PC-Task CPU Test on an 060/100Mhz or do a new video with that? I think it would get some views. Cool that I would find a useful video and it would be yours @Flash951 !

I don't like to post a picture of PCTask running Sysinfo on my 68060@100 because I know the result is totally wrong and messed up.

Here is Sysinfo result on my A4000 from my BB card, MHz is messed up though, 486 CPU is running at 60 MHz.

BTW: A2386SX got a port for a small box that automatically switch monitor between VGA card or Amiga, if that was your issue with BB.

 

[YouKnowWho]

Come on @Flash951...let us see that PC-Task "test result" on your beast Amiga. Please!

 

[Flash951]

Here you go! Totally messed up result. Even the previous posted picture of my A2386SX 486 results is misleading, as the small Sysinfo code probably loops in the 8KB CPU cache, and not from the bottle neck 16 bit system bus on the board. But way more realistic than PCTask results.

Its more like the non-related Sysinfo-program on Amiga when testing an A3640 with 060 adapter. It shows the same speed as a real accelerator with a fast system bus and fast memory access, in reality its about 2/3 as fast do to all the wait states.

 

[YouKnowWho]

Ha ha ha! Thanks!

This is actually informative, as it is funny. I think we can maybe make sense of something here between these two tests.

We have a nice civilized/standard 68060 50Mhz showing 18.9 CPU score on PC-Task and we have your supercharged 68060 at 100Mhz giving us 82.3 PC-Task score.

We have your A2286AT video showing 4.4 CPU score and @Xanxi's A2386SX showing 18.9 CPU score.

Both of these scores are about 4.3x difference, which is an interesting coincidence.

68060 50Mhz is just about 40 Mips and 68060 100Mhz is just about 80 Mips right? Is that 2x or 4x? How is that viewed in terms of CPU performance in real execution time of a task, approximately?

Wonder what PC-Task tweaked to get the results, and if another test can be used for verification. Wasn't this a period in even PC history where these testing apps were gamed and questioned and PC magazines used a bunch of tests to get a more truthful reading?

Meanwhile, this is certainly a reason to get a 68060 going at 100Mhz in your Amiga...you get a 486SX66 free with that upgrade!

 

[Xanxi]

Yes those benchmarks don't necessary relate to real use.
PCTask can do productivity just fine but becomes really slow with graphics even with a lot of RAM dedicated to the video emulation. No sound besides PC beeper but it beeps good.
As for the BB, when you are like me into using games from around 1990-1993, those games which were in competition with the Amiga like Monkey 1 or 2 VGA, Dune (which is better than the Amiga version, Dune II is not or equal in my opinion), all those Sierra games, Wing Commander, etc, there is not a lot of difference to the eyes between 386sx level and 486SLC2. 486SLC2 is necessary for some bigger games, with the limit that Doom on the Amiga 68060 beats easily Doom on the 486SLC2 BB 50 MHz with a good VGA card. That's why i see no point in overclocking such precious hardware and risking destroying it.
The upgraded A2386 is one of the cause of my A4000 reaching very high temperatures inside, useful in winter, limiting in summer.

 

[Flash951]

Yes those benchmarks don't necessary relate to real use.
PCTask can do productivity just fine but becomes really slow with graphics even with a lot of RAM dedicated to the video emulation. No sound besides PC beeper but it beeps good.
As for the BB, when you are like me into using games from around 1990-1993, those games which were in competition with the Amiga like Monkey 1 or 2 VGA, Dune (which is better than the Amiga version, Dune II is not or equal in my opinion), all those Sierra games, Wing Commander, etc, there is not a lot of difference to the eyes between 386sx level and 486SLC2. 486SLC2 is necessary for some bigger games, with the limit that Doom on the Amiga 68060 beats easily Doom on the 486SLC2 BB 50 MHz with a good VGA card. That's why i see no point in overclocking such precious hardware and risking destroying it.
The upgraded A2386 is one of the cause of my A4000 reaching very high temperatures inside, useful in winter, limiting in summer.

I've got a cooling fan on the 486 CPU, no problem with heat. Also fan on the 060 and an upgraded silent fan on psu.

Here is my A2386SX (486) Sysinfo:

 

[Flash951]

Ha ha ha! Thanks!

This is actually informative, as it is funny. I think we can maybe make sense of something here between these two tests.

We have a nice civilized/standard 68060 50Mhz showing 18.9 CPU score on PC-Task and we have your supercharged 68060 at 100Mhz giving us 82.3 PC-Task score.

We have your A2286AT video showing 4.4 CPU score and @Xanxi's A2386SX showing 18.9 CPU score.

Both of these scores are about 4.3x difference, which is an interesting coincidence.

68060 50Mhz is just about 40 Mips and 68060 100Mhz is just about 80 Mips right? Is that 2x or 4x? How is that viewed in terms of CPU performance in real execution time of a task, approximately?

Wonder what PC-Task tweaked to get the results, and if another test can be used for verification. Wasn't this a period in even PC history where these testing apps were gamed and questioned and PC magazines used a bunch of tests to get a more truthful reading?

Meanwhile, this is certainly a reason to get a 68060 going at 100Mhz in your Amiga...you get a 486SX66 free with that upgrade!

You don't get near a 486SX in PCTask, as Ive explained.

You didn't comment my A2386SX Sysinfo result, which are way more realistic.