Jump to content
IGNORED

Atari Falcon: DRAM wait states


jltursan

Recommended Posts

After upgrading my Falcon with a 14MB expansion some time ago, now I'm trying to squeeze the real speed a 60ns SIMM could offer. Following the available documentation online seems an easy operation: the jumper 1 of U46 must be open, this makes the Falcon work with 0-wait states DRAM and it'll run faster than the same machine equipped with a 80ns SIMM and 1-wait state.

 

So good so far, after removing the thin blob of U46-1, I've benchmarked again the Falcon (using the last GemBench 6) and to my dismay I've found that the RAM test run at 77% the standard speed (98%). I guess the reference machine is a Falcon030 with 1-wait state cause my machine gets nearly the 100% when its unmodified; but the new result is totally unexpected, about 25% speed loss.

 

I've checked several SIMMs, all of them 60ns, with the same results. Most of them are EDO RAM but seems to work without glitches so there must be another reason.

 

Has anyone modded the DRAM configuration in his Falcon with good results?

 

 

 

Link to comment
Share on other sites

I did not make tests about RAM on Falcon, but know that RAM access speed depends from used video mode. In higher modes (more res. , colors) there are wait states added for CPU, because video generation bandwith utilizes RAM too much, and can not share it with CPU without added wait states.

Did you try that Gembench test in diverse video modes ?

  • Like 1
Link to comment
Share on other sites

Thanks for the info!, I'm trying to configure desktop as per the machine reference (suposedly a Falcon030 with a 4 colors desktop); but comparing only my own results, it's really strange that the 0-wait state config gives a slower performance value. No changes between tests, only the jumper on/off, 0-wait gives 77% and 1-wait shows the standard performance, about 98%, weird!.

Note that I've not messed with the other jumpers like "16/32 bit video mode" (no idea about its effect) or "TOS ROM wait states" ,all of them are in they factory defaults. I've plans also to change the TOS eeprom to a new faster one; but my first attempt seems that has ruined a AT27C4096, damn PLCC adapter...

 

I'm wondering if my 14MB adapter has somekind of hardwired behaviour or maybe I need to change GemBench6 for another RAM benchmarking tool, I can't believe all the info I've found is plainly wrong!

 

Btw, as I'm also checking HDD performances I'm using right now your HDDAST4 tool, nice app!; so, thanks again for it!

Link to comment
Share on other sites

  • 5 weeks later...

A little update, I've finally been able to get logical benchmarking values for the jumper configuration selected (0 wait for DRAM); but not before setting the ROM wait to 0 and succesfully upgrading TOS IC to a faster one, a 27C4096-85. No idea why but now, setting and unsetting the DRAM wait states gives me some big differences, nearly 20% increase in most tests (ROM has jumped to 119%).

 

As a side note, GemBench is giving me some headaches as it consistenly hangs the machine when running over FreeMint (memory protection?), it works fine when it runs over a plain GEM desktop...

 

All in all, the Falcon feels snappier than ever :-)

Edited by jltursan
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...