User avatar
iljitsch
Website

Posted Tue Aug 10, 2021 3:28 am

SPiFF wrote:
Mon Aug 09, 2021 3:26 pm
Looking around, the TF1260 seems to have good memory speed just the FW is currently not taking full advantage.
Are you sure about the former? This is what I get with my Blizzard:

Code: Select all

BusSpeedTest 0.19 (mlelstv)   Buffer:     262144 Bytes, Alignment: 32768
========================================================================
memtype   addr       op         cycle     calib         bandwidth
fast      $68518000  readw      67.5 ns   normal      29.6 * 10^6 byte/s
fast      $68518000  readl     112.3 ns   normal      35.6 * 10^6 byte/s
fast      $68518000  readm     113.0 ns   normal      35.4 * 10^6 byte/s
fast      $68518000  writew     82.3 ns   normal      24.3 * 10^6 byte/s
fast      $68518000  writel    166.3 ns   normal      24.1 * 10^6 byte/s
fast      $68518000  writem    163.8 ns   normal      24.4 * 10^6 byte/s
chip      $00080000  readw     906.6 ns   normal       2.2 * 10^6 byte/s
chip      $00080000  readl     908.0 ns   normal       4.4 * 10^6 byte/s
chip      $00080000  readm     907.9 ns   normal       4.4 * 10^6 byte/s
chip      $00080000  writew    709.3 ns   normal       2.8 * 10^6 byte/s
chip      $00080000  writel    709.0 ns   normal       5.6 * 10^6 byte/s
chip      $00080000  writem    709.0 ns   normal       5.6 * 10^6 byte/s
(Interesting that reading words is pretty much 60 ns, the memory speed, but writing longs takes just under twice as long.)

Especially my write speeds are three times as fast as yours, which are also worse than on the A3000/25:

Code: Select all

BusSpeedTest 0.19 (mlelstv)   Buffer:     262144 Bytes, Alignment: 32768
========================================================================
memtype   addr       op         cycle     calib         bandwidth
fast      $07A50000  readw     249.7 ns   normal       8.0 * 10^6 byte/s
fast      $07A50000  readl     330.2 ns   normal      12.1 * 10^6 byte/s
fast      $07A50000  readm     302.3 ns   normal      13.2 * 10^6 byte/s
fast      $07A50000  writew    246.8 ns   normal       8.1 * 10^6 byte/s
fast      $07A50000  writel    244.9 ns   normal      16.3 * 10^6 byte/s
fast      $07A50000  writem    219.6 ns   normal      18.2 * 10^6 byte/s
chip      $00018000  readw     817.3 ns   normal       2.4 * 10^6 byte/s
chip      $00018000  readl     812.9 ns   normal       4.9 * 10^6 byte/s
chip      $00018000  readm     636.1 ns   normal       6.3 * 10^6 byte/s
chip      $00018000  writew    574.0 ns   normal       3.5 * 10^6 byte/s
chip      $00018000  writel    573.4 ns   normal       7.0 * 10^6 byte/s
chip      $00018000  writem    577.2 ns   normal       6.9 * 10^6 byte/s
This in line with the Twitter post you linked.

Amazing how much faster chip mem is on the 3000. Note that I ran my chipmem test in PAL 8 color mode after a few tests with other modes, as that makes a significant difference. (Anything 16 colors or less should be fine.)

With a small buffer that fits in the cache our systems perform pretty much identically, one clock cycle per cache read/write, it seems:

Code: Select all

BusSpeedTest 0.19 (mlelstv)   Buffer:       2048 Bytes, Alignment: 32768
========================================================================
memtype   addr       op         cycle     calib         bandwidth
fast      $684F8000  readw      20.3 ns   normal      98.3 * 10^6 byte/s
fast      $684F8000  readl      20.2 ns   normal     198.0 * 10^6 byte/s
fast      $684F8000  readm      20.3 ns   normal     197.1 * 10^6 byte/s
fast      $684F8000  writew     20.6 ns   normal      97.3 * 10^6 byte/s
fast      $684F8000  writel     20.0 ns   normal     199.6 * 10^6 byte/s
fast      $684F8000  writem     20.3 ns   normal     197.2 * 10^6 byte/s
(Rev 5 or 6 68060)
SPiFF wrote:
Fri Jul 30, 2021 11:05 am
It is basically for the overclockers. Nothing for those of us mostly wanting to run at stock speeds.
Hm, I opened the trapdoor to swap out my DIMM to see if that would make a difference with bustest (it didn't) and took a photo of the 68060:
IMG_4568.jpg
So it seems that I have the rev 2, not the rev 5. And it gets pretty hot, even at the stock 50 MHz!

Not sure how much use it is to increase the CPU speed in an A1200 much more, as my impression is that it's the chip mem and the I/O that slows everything down. Although I guess some on-accelerator I/O like with the Vampire will take care of that.

Back to the FPU: I got "flops" from Github to test floating point performance on my 3000 and 1200. I also made the mistake of trying it on my Mac first...

Anyway, the A3000's 68030 can muster about 50 flops with the generic version that I assume is targeted for the 68000. (I used GCC, with -O2 optimizations in all cases. And note to use -DAmiga rather than -DUNIX.) The Blizzard is a bit more than ten times faster at a little over half a megaflop. Compiling for 68020 or 68030 didn't make a noticeable difference (and those two versions are the same size, probably identical.)

Things get more interesting with -m68030 -m68881, so 68030 and FPU. Now the A3000 results shot up to over a quarter megaflop! So that's 4 - 5.5 times faster than without the FPU. On the 68060 this version was also dramatically faster, with 3 - 13 megaflops, so 5 - 20 times faster than the no-FPU version.

So then I compiled for 68040 and 68060, which automatically enables the use of FPU code. Although those two executables weren't identical, the results were identical with the 68030+68881 version. I can claim actually identical because the results for the eight tests also show the size of the error in the calculation, and depending on the type of code used, this differs. 68030+68881, 68040 and 68060 all had the same error.

And now what we've all been waiting for: 68060 but no FPU (-m68060 -msoft-float). Those results are identical to the generic 68000 code: a bit over half a megaflop and at least 3 x slower than a 25 MHz 68030 + 68882. (Using the 68030 + 68881 version.)

Still not sure though what kind of software uses a significant amount of floating point math so you're going to see a difference. But unless the 68LC060 emulates the FPU instructions in software, I'm pretty sure the big issue is going to be having 68030 and up versions of programs assuming the presence of an FPU and then not working.

User avatar
SPiFF

Posted Tue Aug 10, 2021 8:28 pm

iljitsch wrote:
Tue Aug 10, 2021 3:28 am
Note that I ran my chipmem test in PAL 8 color mode after a few tests with other modes, as that makes a significant difference. (Anything 16 colors or less should be fine.)
I ran in 32 colors. There is 16.

Code: Select all

BusSpeedTest 0.19 (mlelstv)   Buffer:     262144 Bytes, Alignment: 32768
========================================================================
memtype   addr       op         cycle     calib         bandwidth
chip      $00040000  readw    1128.6 ns   normal       1.8 * 10^6 byte/s
chip      $00040000  readl    1129.7 ns   normal       3.5 * 10^6 byte/s
chip      $00040000  readm    1130.4 ns   normal       3.5 * 10^6 byte/s
chip      $00040000  writew    887.2 ns   normal       2.3 * 10^6 byte/s
chip      $00040000  writel    888.0 ns   normal       4.5 * 10^6 byte/s
chip      $00040000  writem    887.9 ns   normal       4.5 * 10^6 byte/s
iljitsch wrote:
Tue Aug 10, 2021 3:28 am
Hm, I opened the trapdoor to swap out my DIMM to see if that would make a difference with bustest (it didn't) and took a photo of the 68060:

IMG_4568.jpg

So it seems that I have the rev 2, not the rev 5. And it gets pretty hot, even at the stock 50 MHz!
Yea, those early ones are pretty hot. The "cpu" command on 3.2 reads the revision of the CPU.
iljitsch wrote:
Tue Aug 10, 2021 3:28 am
Still not sure though what kind of software uses a significant amount of floating point math so you're going to see a difference. But unless the 68LC060 emulates the FPU instructions in software, I'm pretty sure the big issue is going to be having 68030 and up versions of programs assuming the presence of an FPU and then not working.
Cool stuff. If you want to share the bins I can run them on the LC060.

Bottom line, if I ever need to run any FPU heavy code I should put my Apollo Turbo 1230 board back in! :mrgreen:

User avatar
iljitsch
Website

Posted Tue Aug 10, 2021 10:02 pm

3.1.4 "cpu" and Whichamiga tell me it's rev 5...
SPiFF wrote:
Tue Aug 10, 2021 8:28 pm
Cool stuff. If you want to share the bins I can run them on the LC060.

Bottom line, if I ever need to run any FPU heavy code I should put my Apollo Turbo 1230 board back in! :mrgreen:
:D

Here's flops with the versions I compiled:

flops-master.zip
(96.15 KiB)

User avatar
SPiFF

Posted Sat Sep 04, 2021 2:54 pm

iljitsch wrote:
Tue Aug 10, 2021 10:02 pm

Here's flops with the versions I compiled:

flops-master.zip
Thanks!
SPiFF wrote: Bottom line, if I ever need to run any FPU heavy code I should put my Apollo Turbo 1230 board back in! :mrgreen:
Er, no. My TurboLC 1230 with 68882 is not much faster than the software float on 060.

Code: Select all

   FLOPS C Program (Double Precision), V2.0 18 Dec 1992

   Module     Error        RunTime      MFLOPS
                            (usec)
     1      5.2023e-10     43.3200      0.3232
     2     -8.8757e-11     24.5600      0.2850
     3     -6.9273e-12     49.6400      0.3425
     4     -1.9554e-11     42.3200      0.3544
     5     -7.4974e-12     77.6800      0.3733
     6     -3.9213e-12     76.6000      0.3786
     7      4.4132e-08     47.2400      0.2540
     8     -5.5261e-12     79.0000      0.3797

   Iterations      =     500000
   NullTime (usec) =     0.2400
   MFLOPS(1)       =     0.3016
   MFLOPS(2)       =     0.3176
   MFLOPS(3)       =     0.3511
   MFLOPS(4)       =     0.3676
Went ahead and run bustest on 020, Apollo 1230 and the TF1260.

Amiga 1200 68020 @ 14Mhz:

Code: Select all

System: 68020   no FastROM (INST: Cache)

Central Processing Unit: MC68020 14.2 MHz
    Floating Point Unit: not available
 Memory Management Unit: not available 
 Max. Chipmem available: 2032 K
 Max. Fastmem available: 0 K

Code: Select all

BusSpeedTest 0.19 (mlelstv)   Buffer:     262144 Bytes, Alignment: 32768
========================================================================
memtype   addr       op         cycle     calib         bandwidth
chip      $00168000  readw     907.9 ns   normal       2.2 * 10^6 byte/s
chip      $00168000  readl     907.4 ns   normal       4.4 * 10^6 byte/s
chip      $00168000  readm     705.9 ns   normal       5.7 * 10^6 byte/s
chip      $00168000  writew    577.3 ns   normal       3.5 * 10^6 byte/s
chip      $00168000  writel    576.1 ns   normal       6.9 * 10^6 byte/s
chip      $00168000  writem    575.5 ns   normal       6.9 * 10^6 byte/s
Apollo Turbo 1230LC w/ 68882 @ 33Mhz, 60ns SIMM:

Code: Select all

System: 68030 68882 68030-MMU FastROM (INST: Cache Burst) (DATA: Cache NoBurst)

Central Processing Unit: MC68030 33.0 MHz
    Floating Point Unit: MC68882
 Memory Management Unit: 68030mmu running
 Max. Chipmem available: 2032 K
 Max. Fastmem available: 8192 K
     Expansion board(s):
2206/9: ELBOX-Computer 1230 (@$00200000 8M)

Code: Select all

BusSpeedTest 0.19 (mlelstv)   Buffer:     262144 Bytes, Alignment: 32768
========================================================================
memtype   addr       op         cycle     calib         bandwidth
fast      $003C0000  readw     187.9 ns   normal      10.6 * 10^6 byte/s
fast      $003C0000  readl     251.7 ns   normal      15.9 * 10^6 byte/s
fast      $003C0000  readm     232.8 ns   normal      17.2 * 10^6 byte/s
fast      $003C0000  writew    189.0 ns   normal      10.6 * 10^6 byte/s
fast      $003C0000  writel    190.4 ns   normal      21.0 * 10^6 byte/s
fast      $003C0000  writem    171.4 ns   normal      23.3 * 10^6 byte/s
chip      $00028000  readw     886.3 ns   normal       2.3 * 10^6 byte/s
chip      $00028000  readl     892.2 ns   normal       4.5 * 10^6 byte/s
chip      $00028000  readm     641.3 ns   normal       6.2 * 10^6 byte/s
chip      $00028000  writew    573.1 ns   normal       3.5 * 10^6 byte/s
chip      $00028000  writel    575.9 ns   normal       6.9 * 10^6 byte/s
chip      $00028000  writem    575.6 ns   normal       6.9 * 10^6 byte/s
TF1260 060LC @ 50Mhz:

Code: Select all

System: 68060  68060-MMU FastROM (INST: Cache Burst) (DATA: Cache CopyBack)
68060 : Rev4 Superscalar Load/StoreBypass StoreBuffer BranchCache 

Central Processing Unit: MC68LC060 42.5 MHz (rev 4)
    Floating Point Unit: not available
 Memory Management Unit: 68060mmu running
 Max. Chipmem available: 2032 K
 Max. Fastmem available: 131072 K
     Expansion board(s):
5080/130: TerribleFire Unknown Prod.ID #130 (@$00E90000 64k)

Code: Select all

BusSpeedTest 0.19 (mlelstv)   Buffer:     262144 Bytes, Alignment: 32768
========================================================================
memtype   addr       op         cycle     calib         bandwidth
fast      $08358000  readw     145.7 ns   normal      13.7 * 10^6 byte/s
fast      $08358000  readl     266.9 ns   normal      15.0 * 10^6 byte/s
fast      $08358000  readm     268.5 ns   normal      14.9 * 10^6 byte/s
fast      $08358000  writew    237.1 ns   normal       8.4 * 10^6 byte/s
fast      $08358000  writel    471.2 ns   normal       8.5 * 10^6 byte/s
fast      $08358000  writem    472.1 ns   normal       8.5 * 10^6 byte/s
chip      $00028000  readw    1128.3 ns   normal       1.8 * 10^6 byte/s
chip      $00028000  readl    1130.0 ns   normal       3.5 * 10^6 byte/s
chip      $00028000  readm    1128.7 ns   normal       3.5 * 10^6 byte/s
chip      $00028000  writew    886.1 ns   normal       2.3 * 10^6 byte/s
chip      $00028000  writel    887.9 ns   normal       4.5 * 10^6 byte/s
chip      $00028000  writem    888.7 ns   normal       4.5 * 10^6 byte/s

User avatar
SPiFF

Posted Sat Sep 04, 2021 2:59 pm

Turns out AmigaKit has a version of the RTC module with Sensors. This will be good if (er, when) I get a Rev.6 CPU to overclock. :mrgreen:

A1200-TF1260_AKSensorModule.jpg
A1200-AKSensorDisplay.jpg

User avatar
A1-X1000
Toronto, Canada
Website

Posted Sun Oct 03, 2021 10:19 pm

it's an awesome 060 accelerator for sure even the way it is now in beta ! I've got a full 060 'running fine at 62.5 MHz without a fan or heatsink on it running for hours (mind you it's in a tower so that makes a big difference for cooling)..though I'll eventually put at least a small heatsink and maybe fan on it one day. Happy I kept my spare 060 when I sold my old Apollo 060 a while back as full 060's are hard to find nowadays and cost an arm & leg ...I love watching 060 specific demos and alot of them require the FPU...I also still love messing around with Lightwave and rendering old objects requires an FPU

I saw a YouTube vid from Chucky (I think) that they're working on a new update to be released soon that will fix the chip/fast ram access and also the internal IDE which should get 6 to 7 MB/s which is damn great compared to the internal MB IDE transfer rate. I'm pretty sure we'll have to mail the cards back to the person who sold you the card to get the update ...

User avatar
A1-X1000
Toronto, Canada
Website

Posted Sun Oct 03, 2021 10:29 pm

SPiFF wrote:
Sat Sep 04, 2021 2:59 pm
Turns out AmigaKit has a version of the RTC module with Sensors. This will be good if (er, when) I get a Rev.6 CPU to overclock. :mrgreen:
Well even if you don't find the hard to find Rev.6 you can overclock any 060 rev from 1 to 5 as well. Mine is a Rev1 and can run at 62.5MHz with ease and I've heard some users Overclocking Rev 4 and 5's to 80 Mhz with cooling. It really depends on the individual CPU too...had a friend that couldn't get a Rev3 higher than 55 MHz without it getting unstable and same guy tested a Rev 1 that ran like a dream for says running Demos at 72 MHz .

User avatar
terriblefire

Posted Wed Oct 13, 2021 2:34 pm

Just a couple of points...

I have a Freescale 68060 Rev6 CPU with a 2013 date code on it.

The 060 either has the FPU in it or not. The TF1260 can use any 060.

There will be a chipram speed fw upgrade in the future.





Return to “Hardware”