Hi, I Don't know if this has been pointed out already, but have you enabled the cache's in your code? I do not have experience with doing DSP in a microblaze, but does anyone else have benchmark figures for such operations? I know that it can be tremendously accelerate using the FSL interface and some dedicated hardware, but for a SW only implementation, someone may have figures faster than one per second! Also, you should check the elf to not only check that it uses the floating point instructions, but also that it does NOT use the emulated double precision insturctions, by searching for the functions that implelement these (I do not know what these functions are called off the top of my head, but you can check this by creating a small program file which explicitly uses doubles, and check that for the function names!) Hope this helps john -----Original Message----- From: owner-microblaze-uclinux@xxxxxxxxxxxxxx on behalf of Abot Botbot Sent: Thu 10/27/2005 2:23 PM To: microblaze-uclinux@xxxxxxxxxxxxxx Subject: RE: [microblaze-uclinux] Jtag Debug Kevin, I know! My whole design team has been grappling around this subject for a month now. We cannot seem to make it go any faster! The performance I'm seeing is as though its running off a 200k Hz clock, which is what the JTAG clk is if I remember correctly. The thing that made me realize that was moving my design from a 50MHz SP3-400 board to the 100MHz ML401 board. Even just from changing the SDRAM bottleneck to a DDR bottleneck, I should have seen an increase in performance. The two designs run at startlingly similar rates. The code I run does 10 FFTs, with the counter you saw timing out the interval necissary to compute those 10 FFTs. That is how I know I get a little better than one FFT per second. I dunno if anyone here is a big EDK user, but that's where my problem probably is. I thoughr the -g option was the only one involved in producing debug symbols and thus would remove any extra software motion caused by the debug hardware if I took it out. Should I build my design from the start with NO debug hardware or software support (one of the early menus in EDK BSB)? I think I'm gonna try that. The other problem is that I havent figured out how to write my configuration and software to flash and then load them at power-on. Due to that limitaiton, my system never runs fully independently, I have to load the code via JTAG and XMD. We gettin there? :D Thanks, Andy >From: "Kevin Chen" <kevin.chen@xxxxxxxxxx> >Reply-To: microblaze-uclinux@xxxxxxxxxxxxxx >To: <microblaze-uclinux@xxxxxxxxxxxxxx> >Subject: RE: [microblaze-uclinux] Jtag Debug >Date: Thu, 27 Oct 2005 11:25:29 +0800 > >Hi, Andy, > >It should not run that slow. There is a timer interrupt connected to the >interrupt controller. Is the interrupt enabled? > >I've done with similar design on ML401 and test the Dhrystones out of >DDR memory. It's quite fast actually. > >kevin > > >-----Original Message----- >From: owner-microblaze-uclinux@xxxxxxxxxxxxxx >[mailto:owner-microblaze-uclinux@xxxxxxxxxxxxxx] On Behalf Of Abot >Botbot >Sent: Wednesday, October 26, 2005 9:27 PM >To: microblaze-uclinux@xxxxxxxxxxxxxx >Subject: [microblaze-uclinux] Jtag Debug > >Hey Friends, > > This might be considered off topic, but I think I'm missing a >fundamantal explanation for the performance that I get. Here's my basic >design flow, which doesn't use uCLinux yet, I was just trying to verify >that the DSP functionality I want is possible on the microblaze... I >download the XMD stub to the board, then run "dow executable.elf" from >XMD and type "run"... >This sets my code in motion, but it seems to be going extremely slow. >It's running entirely out of SRAM, but that shouldn't be the only factor >in the performance I'm seeing... I'm wondering if my whole system is >running over the JTAG clock and putting out debug symbols to the PC, >even though I'm not observing them, and I took the -g option out of the >compiler flags.. So two questions... Am I running in some debug mode >where it is running much slower than it could be... How do I get the >hardware to run independantly, if it is in some debug mode.. After I >finish a few things for my graduate classes (mid terms, blech!) I will >get back on the uCLinux train and get that running on my ML401 board. >Thanks for any help you can give! I'll attatch some files in case anyone >thinks looking at them will help them to understand. > >Thanks, >Andy > >_________________________________________________________________ >Express yourself instantly with MSN Messenger! Download today - it's >FREE! >http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/ > > >___________________________ >microblaze-uclinux mailing list >microblaze-uclinux@xxxxxxxxxxxxxx >Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux >Mailing List Archive : >http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/ > _________________________________________________________________ Express yourself instantly with MSN Messenger! Download today - it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/ ___________________________ microblaze-uclinux mailing list microblaze-uclinux@xxxxxxxxxxxxxx Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/
<<winmail.dat>>