View unanswered posts | View active topics It is currently Mon Jan 21, 2019 7:56 pm



Reply to topic  [ 378 posts ]  Go to page Previous  1 ... 22, 23, 24, 25, 26
 Thor Core / FT64 
Author Message

Joined: Sat Feb 02, 2013 9:40 am
Posts: 740
Location: Canada
I continue to work on both v8 and v7 versions of the core. Switched the DPO branch addressing in v7 back to regular displacement addressing. The DPO addressing wasn’t a bad idea, but it’s also not that good of an idea either. The DPO addressing made it impossible to use compressed branches. Switching back to regular displacements allowed compressed branches, which affected about 5% in compression. Segmentation was also removed from the v7 core, although I keep a copy locally of the segmented core. Paging can adequately provide a virtual address space instead of using segmentation. Privileged access to pages is via keys. The keys are only 10-bit as that’s all the block-ram storage space I was willing to allocate to each page. I’ve seen it recommended that keys be at least 18-bit. There is a keys CSR register that contains up to six ten-bit keys for the app. Every memory access compares all six of the keys in parallel held by the app to the memory page’s key looking for a match. Having the proper key to a page is what allows shared memory. The inverted page table for the system is stored entirely in block-ram memory to allow the fastest processing.
At the moment I’m trying to get a block of pre-initialized data for the standard C library routines copied to an area of ram reserved for the purpose. Many C standard library routines failed to work because duh, the initialization data for the routines wasn’t being setup properly. But the routine hangs at a store word operation.

_________________
Robert Finch http://www.finitron.ca


Thu Jan 10, 2019 7:20 am
Profile WWW

Joined: Sat Feb 02, 2013 9:40 am
Posts: 740
Location: Canada
Received a book on garbage collection a couple of days ago and I’m just reading through it. There’s more to it than I would’ve thought and I’m getting the picture that one solution doesn’t work for everything. So, if GC is included as part of the OS with some hardware support it’ll mean picking and choosing possibly multiple algorithms.

The audio controller in the FT64v7 SoC was generating ack pulses on all writes, not just for itself. This didn’t matter most of the time except for when it did matter.

The IsMem() decoder was missing a default statement to set the output to false, causing some instructions to be interpreted as memory instructions. This caused the core to hang as per the previous post. Another hang fixed.

_________________
Robert Finch http://www.finitron.ca


Fri Jan 18, 2019 6:39 am
Profile WWW

Joined: Tue Dec 11, 2012 8:03 am
Posts: 252
Location: California
robfinch wrote:
Received a book on garbage collection a couple of days ago and I’m just reading through it. There’s more to it than I would’ve thought and I’m getting the picture that one solution doesn’t work for everything. So, if GC is included as part of the OS with some hardware support it’ll mean picking and choosing possibly multiple algorithms.

I'll be anxious to hear more—if I can even understand it.

_________________
http://WilsonMinesCo.com/ lots of 6502 resources


Fri Jan 18, 2019 8:07 pm
Profile WWW
Display posts from previous:  Sort by  
Reply to topic   [ 378 posts ]  Go to page Previous  1 ... 22, 23, 24, 25, 26

Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group
Designed by ST Software