Atari030 wrote:Chris, how does 32Mhz STE compare to a stock MegaSTE? Just curious.
I don't own a MSTE personally, but Rodolphe does
Speed comparisons are tricky because of the cache on the MSTE. While the catch can boost some speeds up, 32mhz will most likely beat it in a lot of tests. I did compare a Falcon to 32mhz, I don't think I posted the results on there, but 32mhz pretty much equalled the falcon in a lot of tests. This is also still at the stock speed for ST-RAM. So once ST-RAM is doubled in speed, those benchmarks will greatly go up in speed.
Falcon on the left, 32mhz STE on the right.
abgdcdfe.png
I also was told my falcon had waitstates, so I removed that and the falcon got a little boost in speed. So display on both machines ended up at about 400%.
CPU figures are a little misleading as the falcon has a FPU so that is why the CPU score is so high, but if you check individual tests, its giving the falcon a run for its money. Now imagine the 020 CPU on the STE with 32bit alt-ram and ROM access, double speeds again. Plus we hope to break the 32mhz barrier on the CPU and aiming for 64mhz. Mix that in with 64mhz alt-ram and getting in the areas of giving the CT60 a run for its money

Well maybe not, but you get the idea

Atari030 wrote:I have near every machine from a 600XL to a pair of Falcons and a TT so boosting an STe personally isn't important. But, I don't mind hacking STFM's. I won't touch my MegaSTE as it is just too risky IMO but I have plenty of spare STFM's.
Lucky you

The STE is going though the 020 route which Rodolphe is working on mostly. I had 32mhz working on the STE in a few hours, and I have been struggling to get 32mhz working at all on my STFM's for months. I'm not sure currently if STE ST-RAM can be boosted in speed as I have not tried it yet, but the STFM is doable, its just a total pain to work with. I wish I had started with the STE, though STE's are harder and more expensive to come by, so its why I started work on STFM's as now and then I tend to kill them

Atari030 wrote:I gotta say I'm impressed at the work you and Rudolph have done. What I find really cool is you two are producing something so powerful and developing it yourselves while countless professional developers back in the day didn't get even close to what you are doing.
Glad someone is enthusiastic about boosters, there doesn't seem many. As Rodolphe said above, back in the early 90s for example, CPU's were expensive, RAM was a huge cost. While there were boosters about, they were expensive. Though also as Rodolphe mentioned, it was speed aswell.
I was really struggling to get the TOS ROM's working at 32mhz with the CPU. The ROM's are 45ns, using "chip enable" but I had to run them with just "output enable" to make them run as 20ns ROM's. With the delay of the GLUE logic, this pushed it up another 10ns or there abouts. Though with the CPU running at 32mhz the time it sets /AS LO to reading DTACK is about 32ns. So the timings were extremely close and its only *just* works. 20ns ROM's are impossible to find (well they are classed as 45ns) Most of the ROMs are 55ns-100ns, They just are not fast enough.
I posted a lot of info about these problems on my site anyway.
http://www.exxoshost.co.uk/atari/last/16mhz/index.htm#DEC92015
You do not have the required permissions to view the files attached to this post.