Jump to content
IGNORED

Atari Disk Speed vs C64 (stock and modded)


bbking67

Recommended Posts

I'm not trying to start an argument or anything, but I am curious... back in the day I remember my buddy's C64 being excruciatingly slow loading disks.  Now the Atari wasn't known for speed either, but the Atari drives always seemed faster than the C64 to me.  Does anyone have any numbers for the stock transfer speeds?

 

The other thing I am interested in are the various speed-ups (from back in the day mainly).  The C64 had the Epyx Fast Load cartridge and the Atari had various drive mods (ICD U.S. Doubler, Happy, etc.).  Was the modded Atari faster than the C64 with Fastload?

 

 

Link to comment
Share on other sites

Stock 1050 transfer speed was 19,200bps.  ICD made a device called the US Doubler (US = Ultra Speed).  It gave the drive true 180kB double density as well as upping the transfer speed to 57,600bps.  I believe the fastest floppy disk for the Atari was the Indus GT using Super SynchMesh.  I know I will be wrong on this, but I believe it was in the range of 68,000bps.

 

I don't have figures on the 1541 speeds, but I will say an incredibly cool feature was being able to use the CPU in the disk drive as a co-processor for the C64.  This was just not possible with the Atari drives, as they had 128bytes of RAM stock and the US doubler gave it 256 bytes of RAM.  The Indus GT was capable of utilizing a full 64kB of RAM and since it also used a Z80 processor as opposed to the Atari drives 6507, it could run CP/M with the Atari acting as the terminal.

  • Like 1
Link to comment
Share on other sites

I used both systems, and in stock form, the C64 was slower, unless you used Atari write with verify.  But using Jiffy Dos or other speeders, the C64 was pretty equivalent to an A8 with Doubler or Happy.  The 1541-II is the C64 drive to have (IMO).  

Link to comment
Share on other sites

Realistically an Atari stock drive will do about 1K per second - to be fair you should calculate a full disk read or at least several tracks.  That averages to a touch over half the actual bitrate used by SIO (which is actually a little under 19.2 k, I don't have the exact figure and it will be a touch slower again for PAL).

 

1541 I think is about 300 bytes per second by default (ref Wiki) which equates to a tad over a quarter the stock Atari speed.

 

Most fastloaders for the 1541 will fairly easily match/beat the stock Atari speed though some are for read operations only (e.g. Epyx Fastload cart).

The quickest fastloaders (from memory I think my brother had a pirated Armalyte that had a mega quick one) are really quick, likely in the same league as an Atari doing 115 kbps.

 

Then you have the hardware mods - I'm reasonably sure there's one that does byte IO for the 1541 which should comfortably beat anything we can do with normal high speed SIO modes (disregarding the recent discussion on doing externally clocked SIO)

 

I've seen a table comparing the most common C64 disk turbo systems but can't remember where.

  • Like 1
Link to comment
Share on other sites

I tend to like to quote the sd2iec firmware wiki page:

 

Speeder C64+1541 C64+sd2iec
relative speed absolute speed relative speed absolute speed
No speeder 1.0x 400 Bytes/Sec 1.6x 650 Bytes/Sec
Turbo Disk 5.7x 2280 Bytes/Sec 12.6x 5050 Bytes/Sec
Jiffy (*1) 6.0x 2400 Bytes/Sec 21.5x 8600 Bytes/Sec
FC3 10.4x 4150 Bytes/Sec 20x 8000 Bytes/Sec
AR6 14.3x 5700 Bytes/Sec (*2) (*2)
SJLOAD (*3)     25x 10000 Bytes/Sec
Proof-of-concept speeder     38x 15300 Bytes/Sec

 

Now 19200 bps would be 2400 bytes/second, so a stock 1050 would equal JiffyDOS. If you obtain 57600 bps that is 7200 bytes/second.

 

The same argument was recently held elsewhere on this forum, comparing Apple II with C64 disk speeds and there were amazing numbers quoted for the A2 under ideal conditions, though in practice those seemed hard to reach.

  • Like 1
  • Thanks 1
Link to comment
Share on other sites

27 minutes ago, carlsson said:

I tend to like to quote the sd2iec firmware wiki page:

 

Speeder C64+1541 C64+sd2iec
relative speed absolute speed relative speed absolute speed
No speeder 1.0x 400 Bytes/Sec 1.6x 650 Bytes/Sec
Turbo Disk 5.7x 2280 Bytes/Sec 12.6x 5050 Bytes/Sec
Jiffy (*1) 6.0x 2400 Bytes/Sec 21.5x 8600 Bytes/Sec
FC3 10.4x 4150 Bytes/Sec 20x 8000 Bytes/Sec
AR6 14.3x 5700 Bytes/Sec (*2) (*2)
SJLOAD (*3)     25x 10000 Bytes/Sec
Proof-of-concept speeder     38x 15300 Bytes/Sec

 

Now 19200 bps would be 2400 bytes/second, so a stock 1050 would equal JiffyDOS. If you obtain 57600 bps that is 7200 bytes/second.

 

The same argument was recently held elsewhere on this forum, comparing Apple II with C64 disk speeds and there were amazing numbers quoted for the A2 under ideal conditions, though in practice those seemed hard to reach.

 

On the Atari side, there will be two (2) concepts involved: "line-rate" speed, and "effective" (-SIO and FileSystem overhead) speed.

 

  1. Line-rate:
    • the Indus/GT (without any HW mode) will be close to 68 Kbps. From there, SIO+FileSystem overhead will need to be subtracted, coming close to 3,750 BYTES/Sec (read-speeds) from magnetic surface.
    • 4,500 BYTES/sec can be achieved when reading from RamCharger Track/Sector Buffer (plug-and-play expansion required).
  2. SIO to PC-based SIO server (like RespeQt):
  • EFFECTIVE speeds of 12,000 BYTES/Sec @ Pokey Divisor 0, including SIO+FileSystem overhead, are the norm.
  • That is in packet-mode transmission offered by PC-Link drivers, all available from SDX command prompt. 
Edited by Faicuai
Link to comment
Share on other sites

Well, Faicuai just mentioned there is a SIO + file system overhead so on the Indus GT with a transfer speed of 68 kbps, the theoretical 6800-8500 bytes/second (depending if every byte counts as 8 or 10 bits) only becomes 3750 bytes/second after those overheads are removed. The numbers in the wiki appear to be effective loading speeds, not theoretical maximum transfer speeds.

 

It can be questioned if JiffyDOS really is slower than FC3 and AR6 but perhaps it never was super fast. I seem to recall that Dolphin DOS with parallel cable was the king on speed but not as common plus that it required a modified drive.

 

Link to comment
Share on other sites

I'm lucky to have Jiffy in my C64 (+SD2IEC). Also have 65XE with SIO2SD. Can't really say I "feel" any difference, both are just fast enough for me.

 

Back in the day that slow FDD speed could be a real problem though, not just inconvenience, for example ruining the Faery Tale Adventure port.

Link to comment
Share on other sites

Instead of synthesizing numbers, actual measurements could be made with a stop watch. Personally I've only got a stock 1050 but could definitely time it vs a 1541 with different software speeders.

 

Stock 1050: Rated for 19200 bps, said to load at around 1024 bytes/second (53% of 1920). Possibly 2.5 times faster than a stock 1541.

Indus GT: Rated for 68000 bps, said to load at around 3750 bytes/second (55% of 6800). Possibly the same speed as a 1541 + FC3.

Link to comment
Share on other sites

2 minutes ago, youxia said:

I'm lucky to have Jiffy in my C64 (+SD2IEC). Also have 65XE with SIO2SD. Can't really say I "feel" any difference, both are just fast enough for me.

 

Back in the day that slow FDD speed could be a real problem though, not just inconvenience, for example ruining the Faery Tale Adventure port.

We didn't know anything else back then, so it wasn't really a problem. Now we can own multiple systems and do comparaisons, but back in the day this was a rare thing. I was a considered a big nerd since I did owned a C64 and an Atari system at the same time, but both were past their prime by then.

Link to comment
Share on other sites

3 hours ago, bbking67 said:

I'm not trying to start an argument or anything, but I am curious... back in the day I remember my buddy's C64 being excruciatingly slow loading disks.  Now the Atari wasn't known for speed either, but the Atari drives always seemed faster than the C64 to me.  Does anyone have any numbers for the stock transfer speeds?

 

The other thing I am interested in are the various speed-ups (from back in the day mainly).  The C64 had the Epyx Fast Load cartridge and the Atari had various drive mods (ICD U.S. Doubler, Happy, etc.).  Was the modded Atari faster than the C64 with Fastload?

 

 

Same

 

BITD, I didn't spend a lot of time benchmarking anything.  I thought the 1050 was a complete godsend compared to a tape drive, that's how I viewed it.  I could access any part of the disk without hitting play or rewind.  

 

Then the byte article came out for the 256k upgrade, and because my dad could so easily do that type of thing, like a week later, we had it installed, and I had a 256k atari, and I used the extra RAM for a ramdisk, to store messages from the BBS.  And that truly made a big speedup.

 

My best friend did own a C64, but he didn't care for disks, he mostly played games.  We argued over whose system was better, but mostly just joking around.  I didn't even use a color screen, I had an amdek amber monitor and went without color or sound.   

 

So while we could argue which system was best for games, he had color, I didn't. 

We could argue which system was best for disk access, but I had a ramdisk, he had only a tape drive. 

 

So arguments tended to be somewhat futile, but it didn't stop us, anyway.

 

I had karateka and pitstop II to demo the atari in games.  Anyway the cool thing about the Atari, is tons of people still had a 2600, so we could argue it was the better Atari, if not the better computer.  

 

We loved to argue, of course we were just as likely to argue over whether "Phi Slamma Jamma" would win the tournament. 

 

 

 

  • Like 1
Link to comment
Share on other sites

2 hours ago, Stephen said:

Stock 1050 transfer speed was 19,200bps.  ICD made a device called the US Doubler (US = Ultra Speed).  It gave the drive true 180kB double density as well as upping the transfer speed to 57,600bps.  I believe the fastest floppy disk for the Atari was the Indus GT using Super SynchMesh.  I know I will be wrong on this, but I believe it was in the range of 68,000bps.

 

I don't have figures on the 1541 speeds, but I will say an incredibly cool feature was being able to use the CPU in the disk drive as a co-processor for the C64.  This was just not possible with the Atari drives, as they had 128bytes of RAM stock and the US doubler gave it 256 bytes of RAM.  The Indus GT was capable of utilizing a full 64kB of RAM and since it also used a Z80 processor as opposed to the Atari drives 6507, it could run CP/M with the Atari acting as the terminal.

So I had two Indus GT drives (which I no longer own) and my units were stock... the syncro-mesh that mine had (non super synchromesh im assuming) wasnt as fast as the US Doubler... that and not much supported it otrher than DOS XL until SpartaDOS X came out.  It was pretty disappointing to me to be honest.  Mind you I possibly just didn't have access to the right mods or software to make it work better.  Over time I ended up preferring the 1050's with US Doubler (which I still own!).  I had no idea that the Indus was capable of such high speed... I'm assuming the 64K upgrade is required for Super Syncro-Mesh?  Or is it just an updated ROM?  I tried so hard to like DOS XL, but I couldnt get into it... SpartaDOS 3.2 was my DOS of choice, but early on SmartDOS was the one I liked with the Indus (no Syncromesh though).

 

Link to comment
Share on other sites

1 hour ago, ivop said:

@carlsson I believe the Speedy upgrade reaches around 80kbps. By your way of calculating, that would mean 10000 Bytes/Sec, similar to SJLOAD.

 

It seems you are forgetting about start and stop-bits ;)

 

 

The Speedy was advertized with 78kBaud, thinking of multiples of 19.200 this should be 76.800 baud (4x), but in reality it is much slower and only a tad faster than a Happy drive. I somehow doubt that the Speedy High Speed Sector Copy (or HSS Copy II), which was part of the firmware, really reaches the advertized 96k Baud. Maybe someone can measure the mini Speedy's speed a) in normal mode (a bit faster than 19k2 baud, which would be slow mode), b) in ultraspeed mode and c) with HSS Copy or HSS Copy II.

 

Think the german Turbo 1050 is comparable to an Indus GT drive, it reaches 68k2 Baud with special sector interleave (known as turbo-format). I once tested both mini Speedy 1050 and Turbo 1050 and when both were set to highest speed*, the Turbo was somewhat faster in reading and writing the same disk than a Speedy; but the Speedy was so much faster in formatting the disk (Turbo 1050 was/is as slow as a standard 1050 drive when formatting), that it won the whole diskcopy speedtest. * Did not use HSS Copy or HSS Copy II with the mini Speedy drive, instead used 1) Diskcopy (from TurboDOS XL/XE) for both Speedy and Turbo and also 2a) Sectorcopy 1.5 for Speedy vs. 2b) Trackcopy for Turbo. 

 

The Super Speedy, Mega Speedy and mini Super Speedy however do have so much RAM onboard, that they can copy a 90k/130k/180k disk within a few seconds internally (and the special Super Speedy Copy program also recognizes if a disk is already correctly formatted, so it will not format it again).

 

Link to comment
Share on other sites

27 minutes ago, bbking67 said:

So I had two Indus GT drives (which I no longer own) and my units were stock... the syncro-mesh that mine had (non super synchromesh im assuming) wasnt as fast as the US Doubler... that and not much supported it otrher than DOS XL until SpartaDOS X came out.  It was pretty disappointing to me to be honest.  Mind you I possibly just didn't have access to the right mods or software to make it work better.  Over time I ended up preferring the 1050's with US Doubler (which I still own!).  I had no idea that the Indus was capable of such high speed... I'm assuming the 64K upgrade is required for Super Syncro-Mesh?  Or is it just an updated ROM?  I tried so hard to like DOS XL, but I couldnt get into it... SpartaDOS 3.2 was my DOS of choice, but early on SmartDOS was the one I liked with the Indus (no Syncromesh though).

 

I don't know honestly - I never was able to get high speed working with my Indus.  I never had it back in the day when I was using Sparta DOS 3.x.  I now try to exclusively use SDX, and the only thing I use the Indus for is CP/M.

Link to comment
Share on other sites

It's been a while, but I recall that the C64 drives were supposed to be much faster than they were but there was a coding issue/bug/chip issue  that had to be worked around at the last minute before production.  It might be in one of the Bagnall books, or one on MOS.

Link to comment
Share on other sites

Internallly stock C64 running dual emulated 1541's with 40k RAMBoards via the 1541 UII+ using the standard IEC serial connector (note, the RAMBords make no difference to transfer speeds, emulated 1541's are perfectly cycle accurate).

 

1. S-JiffyDOS ROM's on the 1541, standard JiffyDOS 6.01 on the C64:

 

WqXHjHD.jpg

 

buCa3kw.jpg

 

Vs the C64 once again running JiffyDOS 6.01 but with the 1541 running standard JiffyDOS ROMs and 40k RAMBoards:

 

bGJKVEq.jpg

 

BxuiIV4.jpg

 

Now this is impressive, be aware that naturally we're including overheads here and not just sprouting on about theoretical maximums. This is running the Hyperspeed kernel via the parallel bus of the cartridge port using DMA as intended by the Commodore engineers. No screen blanking needed, no cheating in any way whatsoever:

 

aiY4tgt.jpg

 

So basically, I interpret those figures as the C64 and the A8 trading blows in terms of storage throughput. For ancient machines such speeds are impressive. Both amazing platforms.

 

Edited by Mazzspeed
  • Like 5
Link to comment
Share on other sites

Here is the timing for a stock 1050, reading then writing an entire SD disk (92,160 bytes).  Times by stopwatch. Copy 130XE sector copier used to read the entire disk into memory.  NTSC system.  I'll try to add Happy, US, etc, if no one does it before. BTW, this throughput is what Rybags posted above in #4.

 

Read time: 85 seconds.  Throughput 1084 Bps.

Write time 85 seconds. Throughput 1084 Bps. (No verify)

 

Happy 1050 (No Warp)  The Happy 810 was the first Atari "speeder."  Unfortunately I no longer have any 810's.

Read Time: 65 seconds.  Throughput 1418 Bps.

Write Time: 85 seconds.  Throughput 1084 Bps.

 

Ultra Speed 1050 

Read Time: 47 seconds.  Throughput 1960 Bps

Write Time: 48 seconds.  Throughput 1920 Bps

 

  • Like 1
Link to comment
Share on other sites

19 hours ago, Tuxon86 said:

We didn't know anything else back then, so it wasn't really a problem.

In some cases, like the Faery Tale I mentioned, it was. It made the game (a huge cRPG ported from Amiga) practically unplayable for somebody without a fastloader. It's what the actual reviewers said back then.

 

Otherwise it was just an inconvenience: I sure as hell would prefer a "slow" 1541, even without any speed up thingies, to the tape deck I was stuck with ?

Link to comment
Share on other sites

I had Dolphin Dos with the cable back in the day, yep it was incredibly fast, the sector copy was like lightning but in truth it was as much trouble as it was worth with protected disks ala the Happy issue.. I'm pleased with both systems adjusted loading abilities especially compared to an untouched non fast loading 1541 which I seem to remember being quoted as slower than a fast loading tape at times..

Link to comment
Share on other sites

11 hours ago, kheller2 said:

the C64 drives were supposed to be much faster than they were but there was a coding issue/bug/chip issue

Yes, a problem with the shift register in the 6522 VIA found in the drive. It might be so that the 6526 CIA corrects this, but since there were existing 1540 drives on the market (that eventually could be upgraded to 1541), Commodore decided to use the slower speed instead of making a new model that as well would not be backwards compatible with eventual VIC-20 users who wanted a floppy drive. On the PET side it probably never was a problem since the IEEE uses parallel data.

 

So sure, if Commodore had thrown VIC owners "under the bus" when they released the C64, the 1541 probably could have been faster from the factory. I don't know how well 1050 and later drives work with e.g. an Atari 400, perhaps there never were any compatibility issues around that?

Link to comment
Share on other sites

35 minutes ago, Larry said:

Stock 1050

Read time: 85 seconds.  Throughput 1084 Bps.

Write time 85 seconds. Throughput 1084 Bps. (No verify)

 

Happy 1050 (No Warp)

Read Time: 65 seconds.  Throughput 1418 Bps.

Write Time: 85 seconds.  Throughput 1084 Bps.

Thanks for that. It resembles what Rybags wrote above, that a stock 1050 would do around 1 kB per second, and just above the calculations I did (using a value of 1024 bytes/second). So combined with the measurements Mazzspeed did simultaneously, it really looks like stock 1050 = 2.5 times faster than stock 1541. That is quite a bit faster, but not 5X or 10X.

Edited by carlsson
Link to comment
Share on other sites

49 minutes ago, carlsson said:

Yes, a problem with the shift register in the 6522 VIA found in the drive. It might be so that the 6526 CIA corrects this, but since there were existing 1540 drives on the market (that eventually could be upgraded to 1541), Commodore decided to use the slower speed instead of making a new model that as well would not be backwards compatible with eventual VIC-20 users who wanted a floppy drive. On the PET side it probably never was a problem since the IEEE uses parallel data.

 

So sure, if Commodore had thrown VIC owners "under the bus" when they released the C64, the 1541 probably could have been faster from the factory. I don't know how well 1050 and later drives work with e.g. an Atari 400, perhaps there never were any compatibility issues around that?

It was a little more than that.

 

The original design was intended to make full use of the shift registers implemented in the 6522 VIA, and designs were sent off to the PCB fab with all the interconnects needed to make it happen. When the finished drive made it's way back to the engineers responsible for the improved serial IEC implementation they found all the new traces on the PCB had been removed making the 1541 actually slower than the 1540 as connected to the VIC20!

 

The excuse from the manufacturing department?

 

The PCB didn't fit in the old 1540 case that was being reused for the 1541 so the improvements had to go - Needless to say the engineers were really pissed!

 

Luckily the community examined the implementation, and now a 1541 running JiffyDOS is actually faster than the unobtanium and very rare (and desireable) SFD1001 FDD with the parallel IEEE-488 connection. It just can't hold that magic 1MB of data the SFD1001 could hold.

 

51 minutes ago, Mclaneinc said:

I had Dolphin Dos with the cable back in the day, yep it was incredibly fast, the sector copy was like lightning but in truth it was as much trouble as it was worth with protected disks ala the Happy issue.. I'm pleased with both systems adjusted loading abilities especially compared to an untouched non fast loading 1541 which I seem to remember being quoted as slower than a fast loading tape at times..

The Ultimate 64 can implement Dolphin DOS as well as the parallel connection virtually in FPGA via the menu, very cool stuff and totally accurate.

Edited by Mazzspeed
  • Like 1
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...