Safe to "under" clock a Cyberstorm 060 MK-2? - Page 2

trixster · 20 May 2020, 17:32

Would be interesting to get the speed of your CSPPC!

trixster · 20 May 2020, 17:34

Quote:

Originally Posted by DarrenHD

Yes and MK3/CS PPC is faster still. Matze's TK-60 is insanely fast have you seen his bustest results @ 100Mhz? like 80MB...

Yup! I’m going to get my name on the list for one of those is there’s a group buy

I have a 48mhz crystal on the way to try the WE at 96mhz

Kin Hell · 20 May 2020, 17:59

So Slightly slower for it Darren.

Very interesting to See those Warp Speeds trixter. - 80Mhz 060 conversion, yeah?

trixster · 20 May 2020, 19:21

Yeah, one of those nifty 040-060 adapters you see people putting on A3640s. There’s a very good old thread on here from 2013 called Trans-Warpengine which shows the details

Kin Hell · 20 May 2020, 19:25

Quote:

Originally Posted by trixster

Yeah, one of those nifty 040-060 adapters you see people putting on A3640s. There’s a very good old thread on here from 2013 called Trans-Warpengine which shows the details

Aye. I remember my Warp Engine card being much much faster than CS MKII 060. The Desktop @ the same Res was scarily Faaast!

It was an Amazing Card for an 040.

DarrenHD · 20 May 2020, 22:03

Quote:

Originally Posted by trixster

Would be interesting to get the speed of your CSPPC!

That can be arranged tonight

hehe

DarrenHD · 21 May 2020, 06:59

Enabled 60ns ram on cs ppc...quite a bit faster.

trixster · 21 May 2020, 07:19

What speed is the csppc running at?

DarrenHD · 21 May 2020, 14:15

Quote:

Originally Posted by trixster

What speed is the csppc running at?

50MHz. 233 for the PPC.

trixster · 21 May 2020, 14:58

Interesting, thanks.

So we can estimate that these would be the results for a WE running it's ram at 50mhz:

readw 49.125
readl 56.125
readm 56.25
writew 35.25
writel 35.125
writem 35.625

So WE a bit faster at read and a little slower at write

DarrenHD · 21 May 2020, 16:19

Quote:

Originally Posted by trixster

Interesting, thanks.

So we can estimate that these would be the results for a WE running it's ram at 50mhz:

readw 49.125
readl 56.125
readm 56.25
writew 35.25
writel 35.125
writem 35.625

So WE a bit faster at read and a little slower at write

Yes the WE is/was a great (the best) 040 accelerator. Do you know if you fit it with an 060 if it affects memory speed? Would be interesting to see that result. In any case Matze's TK-060 will be awesome, but too bad it doesn't have on-board SCSI, on the CS PPC what really helps to speed up the system too when you have 30MB/sec...

trixster · 21 May 2020, 16:35

not sure, it's a hassle to dismantle the AA3000+ to fit an 040 to it to test! My test is with an 060 at 80mhz (so ram at the 40mhz crystal speed), so maybe the ram scores I'm seeing are a little higher (or lower?) than a 40mhz 040 would give.

I'm actually still using the AA3000+'s onboard SCSI, so only 3.5mb/s, but it's good enough for me at the moment. I havent tried the WE's scsi just yet.

DarrenHD · 21 May 2020, 21:48

Quote:

Originally Posted by trixster

not sure, it's a hassle to dismantle the AA3000+ to fit an 040 to it to test! My test is with an 060 at 80mhz (so ram at the 40mhz crystal speed), so maybe the ram scores I'm seeing are a little higher (or lower?) than a 40mhz 040 would give.

I'm actually still using the AA3000+'s onboard SCSI, so only 3.5mb/s, but it's good enough for me at the moment. I havent tried the WE's scsi just yet.

So you have the Warp Engine in your AA3000+ (in an A3000 case?) It fits ok, or do you have the Warp Engine 3000 version with angled simm sockets?

I plan on putting Matze's TK-060 in my A33000+ (if it ever arrives) into an A3000 case. I am hoping to get more than 3.5MB/sec out of the onboard SCSI with the TK-060 installed. I get 3.3MB/sec now with a C= 3640 in my existing A3000. Do you know what the max speed is for the A3000 onboard SCSI? I have the WD -08 chip.

trixster · 21 May 2020, 22:52

Yeah, angled simm slots and very low profile 040-060 adapter

I think i read somewhere that 3.5mb/s is all you can expect from A3000 scsi. I’m using the AMD AM33C93A scsi chip and an Acard 7720UW which I believe is pretty quick

3mb/a is what I’m seeing too in my A3000 which has an A3640

DarrenHD · 25 May 2021, 15:24

So, and update on my 100MHz Cyberstorm MK2 060...

I'm having some trouble with SCSI disk access, specifically read and write errors, and sometimes HDToolbox locking up. I am not sure exactly what it is, but it may be that the 100MHz hack is not 100% compatible, at least with my board and CPU.

I have tried different SCSI drives and drives connected via an ACard...and I am getting the same errors. Most of the time it is only some write errors, reading is fine. So your system appears to be fine, but when you do lots of writes (like copy a bunch of files) it will usually give a write error.

I am contemplating to go back to the stock 50MHz configuration to hopefully make the SCSI reliable again. While certainly faster for pure CPU operations, the 50Mhz memory is sort of a bottle neck.

Damion · 25 May 2021, 19:04

Quote:

Originally Posted by DarrenHD

So, and update on my 100MHz Cyberstorm MK2 060...

I'm having some trouble with SCSI disk access, specifically read and write errors, and sometimes HDToolbox locking up. I am not sure exactly what it is, but it may be that the 100MHz hack is not 100% compatible, at least with my board and CPU.

I have tried different SCSI drives and drives connected via an ACard...and I am getting the same errors. Most of the time it is only some write errors, reading is fine. So your system appears to be fine, but when you do lots of writes (like copy a bunch of files) it will usually give a write error.

I am contemplating to go back to the stock 50MHz configuration to hopefully make the SCSI reliable again. While certainly faster for pure CPU operations, the 50Mhz memory is sort of a bottle neck.

FWIW, I tried this briefly a while ago and quickly reverted back to the 1/1 setup. Everything "seemed" OK, with CPU tasks naturally a bit snappier. First issue was I noticed that some demos acted funny at 100MHz. Worse, the AmigaMark FPU benchmark would fail 2/5 runs or so. (Sidenote: I had always used Quake for checking FPU stability, but the AmigaMark benchmark appears to be much better/quicker for this purpose.) It's likely the CPU needed a voltage bump, but it didn't seem worth the risk or hassle.

Bummer to lose the CyberSCSI, but IMHO 66MHz with a Fastlane/A4091 still appears to be the best overall compromise with the MK2.

DarrenHD · 25 May 2021, 19:59

Quote:

Originally Posted by Damion

FWIW, I tried this briefly a while ago and quickly reverted back to the 1/1 setup. Everything "seemed" OK, with CPU tasks naturally a bit snappier. First issue was I noticed that some demos acted funny at 100MHz. Worse, the AmigaMark FPU benchmark would fail 2/5 runs or so. (Sidenote: I had always used Quake for checking FPU stability, but the AmigaMark benchmark appears to be much better/quicker for this purpose.) It's likely the CPU needed a voltage bump, but it didn't seem worth the risk or hassle.

Bummer to lose the CyberSCSI, but IMHO 66MHz with a Fastlane/A4091 still appears to be the best overall compromise with the MK2.

Yes for sure. I don't have a 4091 - just the MK2 scsi module...but yeah, I'll be going back to 50Mhz on this card. If I manage to track down a 4091 I can go up to 66Mhz on it.

DarrenHD · 26 May 2021, 15:09

So, I went back to 50MHz, non overclocked last night. Still getting weird write errors, read errors, and sometimes checksum errors. I tried two different hard drives, playing with file systems, direct scsi on/off, mask, maxtransfer, etc.

the funny thing is one of my drives tested does not give errors - that one has a couple of SFS partitions and SFS also in the RDB (along with the new FastfileSystem from 3.2). The other drives are just using FFS. If I hook up these drives to another A4000 on the ide interface they don't give errors.

Tried a different SCSI cable as well....it's really strange. Have to try and figure this out...

DarrenHD · 27 May 2021, 15:14

OK, so I managed to find a combination that appears to be working - no more read/checksum/write errors on the disk. It is very strange. I tested 4 different drives

1) one real SCSI disk (old Quantum) directly on the SCSI cable: errors
2) 150GB WD IDE disk connected to an ACard: errors
3) 60GB Quantum IDE disk connected to an ACard: no errors
4) 32GB Sandisk SD card connected via IDE/CF adapter to the ACard: no errors

It did not matter the file system - at least FFS and PFS both generated errors in 1) and 2) above.

So I have left in option 4) the 32GB Sandisk SD card on the IDE/CF adatper connected via the ACard.

Now the question is...since I also had the errors at 50MHz perhaps it wasn't the overclock as the reason for the errors...so should I go back to 100MHz ?

It's a pain to move the jumpers (very tiny and requires soldering) and one of the pads is already pretty flakey.

Darren

nogginthenog · 27 May 2021, 19:08

Are you sure it's not a SCSI termination issue? The CyberSCSI MK-II external module provides termination.

Pictured here (4th image)
https://bigbookofamigahardware.com/b...t.aspx?id=1230

Stick it on the end of your chain.

21 May 2020, 06:59	#27
DarrenHD Registered User Join Date: Aug 2008 Location: London / Canada Posts: 781	Enabled 60ns ram on cs ppc...quite a bit faster. Attached Thumbnails

21 May 2020, 14:58	#30
trixster Guru Meditating Join Date: Jun 2014 Location: England Posts: 2,339	Interesting, thanks. So we can estimate that these would be the results for a WE running it's ram at 50mhz: readw 49.125 readl 56.125 readm 56.25 writew 35.25 writel 35.125 writem 35.625 So WE a bit faster at read and a little slower at write Last edited by trixster; 21 May 2020 at 15:25.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
"Diabolik" & "Dylan Dog" & "Tex" & "Time Runners" series	DamienD	request.Old Rare Games	20	21 July 2022 16:58
"Voices8" 8 Channel Soundtracker "DemoSongI" song - "This is the Amiga with 8 Voices"	DemosongIHunter	request.Music	45	23 May 2022 20:07
A2000 clock having "problems"	system11	support.Hardware	7	10 November 2017 02:53
"Internal clock not functioning" error	Supamax	support.WinUAE	2	28 January 2015 09:51
"Bottleneck" inscription on the CyberStorm MK1 accelerator...	keropi	Hardware pics	17	07 December 2006 20:54

20 May 2020, 17:32	#21
trixster Guru Meditating Join Date: Jun 2014 Location: England Posts: 2,339	Would be interesting to get the speed of your CSPPC!

20 May 2020, 17:59	#23
Kin Hell 0ld0r Git Join Date: Mar 2009 Location: Cornwall, UK Posts: 1,581	So Slightly slower for it Darren. Very interesting to See those Warp Speeds trixter. - 80Mhz 060 conversion, yeah?

20 May 2020, 19:21	#24
trixster Guru Meditating Join Date: Jun 2014 Location: England Posts: 2,339	Yeah, one of those nifty 040-060 adapters you see people putting on A3640s. There’s a very good old thread on here from 2013 called Trans-Warpengine which shows the details

21 May 2020, 07:19	#28
trixster Guru Meditating Join Date: Jun 2014 Location: England Posts: 2,339	What speed is the csppc running at?

21 May 2020, 16:35	#32
trixster Guru Meditating Join Date: Jun 2014 Location: England Posts: 2,339	not sure, it's a hassle to dismantle the AA3000+ to fit an 040 to it to test! My test is with an 060 at 80mhz (so ram at the 40mhz crystal speed), so maybe the ram scores I'm seeing are a little higher (or lower?) than a 40mhz 040 would give. I'm actually still using the AA3000+'s onboard SCSI, so only 3.5mb/s, but it's good enough for me at the moment. I havent tried the WE's scsi just yet.

21 May 2020, 22:52	#34
trixster Guru Meditating Join Date: Jun 2014 Location: England Posts: 2,339	Yeah, angled simm slots and very low profile 040-060 adapter I think i read somewhere that 3.5mb/s is all you can expect from A3000 scsi. I’m using the AMD AM33C93A scsi chip and an Acard 7720UW which I believe is pretty quick 3mb/a is what I’m seeing too in my A3000 which has an A3640

25 May 2021, 15:24	#35
DarrenHD Registered User Join Date: Aug 2008 Location: London / Canada Posts: 781	So, and update on my 100MHz Cyberstorm MK2 060... I'm having some trouble with SCSI disk access, specifically read and write errors, and sometimes HDToolbox locking up. I am not sure exactly what it is, but it may be that the 100MHz hack is not 100% compatible, at least with my board and CPU. I have tried different SCSI drives and drives connected via an ACard...and I am getting the same errors. Most of the time it is only some write errors, reading is fine. So your system appears to be fine, but when you do lots of writes (like copy a bunch of files) it will usually give a write error. I am contemplating to go back to the stock 50MHz configuration to hopefully make the SCSI reliable again. While certainly faster for pure CPU operations, the 50Mhz memory is sort of a bottle neck.

26 May 2021, 15:09	#38
DarrenHD Registered User Join Date: Aug 2008 Location: London / Canada Posts: 781	So, I went back to 50MHz, non overclocked last night. Still getting weird write errors, read errors, and sometimes checksum errors. I tried two different hard drives, playing with file systems, direct scsi on/off, mask, maxtransfer, etc. the funny thing is one of my drives tested does not give errors - that one has a couple of SFS partitions and SFS also in the RDB (along with the new FastfileSystem from 3.2). The other drives are just using FFS. If I hook up these drives to another A4000 on the ide interface they don't give errors. Tried a different SCSI cable as well....it's really strange. Have to try and figure this out...

27 May 2021, 15:14	#39
DarrenHD Registered User Join Date: Aug 2008 Location: London / Canada Posts: 781	OK, so I managed to find a combination that appears to be working - no more read/checksum/write errors on the disk. It is very strange. I tested 4 different drives 1) one real SCSI disk (old Quantum) directly on the SCSI cable: errors 2) 150GB WD IDE disk connected to an ACard: errors 3) 60GB Quantum IDE disk connected to an ACard: no errors 4) 32GB Sandisk SD card connected via IDE/CF adapter to the ACard: no errors It did not matter the file system - at least FFS and PFS both generated errors in 1) and 2) above. So I have left in option 4) the 32GB Sandisk SD card on the IDE/CF adatper connected via the ACard. Now the question is...since I also had the errors at 50MHz perhaps it wasn't the overclock as the reason for the errors...so should I go back to 100MHz ? It's a pain to move the jumpers (very tiny and requires soldering) and one of the pads is already pretty flakey. Darren

27 May 2021, 19:08	#40
nogginthenog Amigan Join Date: Feb 2012 Location: London Posts: 1,311	Are you sure it's not a SCSI termination issue? The CyberSCSI MK-II external module provides termination. Pictured here (4th image) https://bigbookofamigahardware.com/b...t.aspx?id=1230 Stick it on the end of your chain.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)