search

taijusti / sleep_apnea

0 stars 0 forks source link

K engine interface incompatible #1

Closed taijusti closed 9 years ago

taijusti commented 9 years ago

K engine interface needs to be made compatible with device.c. device_tb.c should pass.

taijusti commented 9 years ago

i can change the directives to in E and KKT pipelines to pipeline II=1 if you don't wanna.

ghost commented 9 years ago

No its okay I want to do them

On Thu, Apr 23, 2015 at 7:00 PM, Justin notifications@github.com wrote:

i can change the directives to in E and KKT pipelines to pipeline II=1 if you don't wanna.

— Reply to this email directly or view it on GitHub https://github.com/taijusti/sleep_apnea/issues/1#issuecomment-95740817.

taijusti commented 9 years ago

ok

taijusti commented 9 years ago

should this be closed? btw, the SW implementation gives different results than the HW implementation i think? i'm trying to get device_tb.c to work right now to test my stuff.

ghost commented 9 years ago

it is compatible now, but I also got co/simulation failure when I tried c against RTL simulation of the device_tb.c. However, the test bench for the kernel passes the co/simulation. Did the co-simulation of device_tb.c pass before (with the dummy kernel)?

On Sat, Apr 25, 2015 at 8:41 PM, Justin notifications@github.com wrote:

should this be closed? btw, the SW implementation gives different results than the HW implementation i think? i'm trying to get device_tb.c to work right now to test my stuff.

— Reply to this email directly or view it on GitHub https://github.com/taijusti/sleep_apnea/issues/1#issuecomment-96299228.

taijusti commented 9 years ago

yup, it passed with the dummy kernel, its because sw_k() inside of device_tb needs to be replaced with the updated one. sw_k() computes the kernel value of one pair of points. could you update this?

ghost commented 9 years ago

I update device_tb.c locally, but it still doesn't pass co-simulation. The mismatch is in the E values. I will check what is wrong on Tuesday.

On Sun, Apr 26, 2015 at 3:44 PM, Justin notifications@github.com wrote:

yup, it passed with the dummy kernel, its because sw_k() inside of device_tb needs to be replaced with the updated one. sw_k() computes the kernel value of one pair of points. could you update this?

— Reply to this email directly or view it on GitHub https://github.com/taijusti/sleep_apnea/issues/1#issuecomment-96426195.

taijusti commented 9 years ago

does it pass c-simulation?

ghost commented 9 years ago

now it passes c-simulation, I committed the changes.

On Mon, Apr 27, 2015 at 1:31 AM, Justin notifications@github.com wrote:

does it pass c-simulation?

— Reply to this email directly or view it on GitHub https://github.com/taijusti/sleep_apnea/issues/1#issuecomment-96512328.

taijusti commented 9 years ago

nice! should we close issue?

ghost commented 9 years ago

yes I think we should. Did you try it ?

On Mon, Apr 27, 2015 at 10:53 AM, Justin notifications@github.com wrote:

nice! should we close issue?

— Reply to this email directly or view it on GitHub https://github.com/taijusti/sleep_apnea/issues/1#issuecomment-96684509.

taijusti commented 9 years ago

c-sim passed. i killed co-sim after 5 minutes or so--i'm guessing it doesn't like expf. should be ok though. closing. just re-open if you have something.