#
#asahi-gpu · 2021-05-10 18:33
azonenbergwatching memory accesses via jtag or something while calling various APIs?
#asahi-gpu · 2021-05-10 18:33
azonenbergor is dynamic analysis fair game?
#asahi-gpu · 2021-05-10 18:32
azonenbergbut judging by the topic that's not the case
#asahi-gpu · 2021-05-10 18:32
azonenbergi assumed most of this kind of stuff had to be done by finding someone in a suitably friendly jurisdiction to reverse the relevant tidbits then write a clean-room spec
#asahi-gpu · 2021-05-10 18:31
azonenbergpoking registers at random seems likely to end badly and not tell you very much
#asahi-gpu · 2021-05-10 18:31
azonenbergOut of curiosity if you're not reversing any apple driver code or decapping the chip, how are you figuring out anything about the register interface?
#asahi-gpu · 2021-05-10 13:15
bloomcompute shaders will happen wen they happen, easier tha 3D but less fun :-p
#asahi-gpu · 2021-05-10 10:10
azonenberglong term a community opengl 4.x implementation that runs under osx would be nice :p
#asahi-gpu · 2021-05-10 10:10
azonenbergHence being interested in potential for it to work on M1 at least under linux
#asahi-gpu · 2021-05-10 10:10
azonenbergAnyway, so right now it won't run on any apple platform unless it's an x86 that you dual boot linux on
#asahi-gpu · 2021-05-10 10:06
marcanah, neat
#asahi-gpu · 2021-05-10 10:05
azonenbergsignal sources (channels or imported files), sinks (displayed plots), and processing blocks between them in an arbitrary DAG
#asahi-gpu · 2021-05-10 10:05
azonenbergglscopeclient works a lot like gnuradio or VTK/ITK
#asahi-gpu · 2021-05-10 10:05
azonenbergbut it's just another DSP block you can put in the pipeline
#asahi-gpu · 2021-05-10 10:04
azonenbergand then render the upsampled waveform with the shader
#asahi-gpu · 2021-05-10 10:04
azonenbergNot in the shader. there's a sinc filter you can apply to a waveform with arbitrary upsample rate if you want it
#asahi-gpu · 2021-05-10 10:01
marcandoes it do sinc interpolation? :>
#asahi-gpu · 2021-05-10 10:00
azonenbergi integrate in shared memory using a couple of threads, then splat out to global memory at the very end since that's slow
#asahi-gpu · 2021-05-10 10:00
azonenbergfor y=min to max waveform[x][y] += alpha
#asahi-gpu · 2021-05-10 10:00
azonenbergfind min/max Y value, interpolating if the segment from that sample to the previous/next crosses outside the pixel
#asahi-gpu · 2021-05-10 09:59
azonenbergwithin each thread: loop over all samples whose timestamp is in that pixel
#asahi-gpu · 2021-05-10 09:59
azonenbergspawn one compute thread for each pixel along the x axis
#asahi-gpu · 2021-05-10 09:59
azonenbergThe waveform rendering itself is basically
#asahi-gpu · 2021-05-10 09:59
azonenbergthird is Cairo again for axis labels, protocol decode annotations, and cursors
#asahi-gpu · 2021-05-10 09:58
azonenbergsecond is the compute shader generated waveform
#asahi-gpu · 2021-05-10 09:58
azonenbergFirst one is Cairo for the grid, background gradient, etc
#asahi-gpu · 2021-05-10 09:58
azonenbergall with a single texture and very simple shader
#asahi-gpu · 2021-05-10 09:58
azonenbergThe glscopeclient renderer displays a total of three fullscreen "quads" of two triangles per render pass
#asahi-gpu · 2021-05-10 09:57
marcanremember this gpu is TBDR; I wonder if that affects your optimizations
#asahi-gpu · 2021-05-10 09:55
azonenbergThe output is a monochrome fp32 texture that i then splat onto a full-viewport pair of triangles and use a fragment shader to tone map to colors
#asahi-gpu · 2021-05-10 09:54
azonenbergnamely that points are sorted in strict left to right order and in many cases at regular intervals (I have optimizations i can ifdef for that to build two different versions of the shader)
#asahi-gpu · 2021-05-10 09:54
azonenbergthe glscopeclient renderer is basically a software rasterizer in compute shaders that takes advantage of the specialized characteristics of waveform data
#asahi-gpu · 2021-05-10 09:53
marcanbut my uneducated guess is that computer shaders are *easier* than 3D
#asahi-gpu · 2021-05-10 09:53
marcancc bloom (I think she's asleep though)
#asahi-gpu · 2021-05-10 09:53
azonenbergright now that's not possible
#asahi-gpu · 2021-05-10 09:52
azonenbergNot an immediate priority but i have a friend getting a M1 mac who is interested in using glscopeclient on it
#asahi-gpu · 2021-05-10 09:52
azonenbergWondering what the status of OpenGL is. Specifically how plausible it is to get GL_ARB_compute_shader in mesa working on the M1 GPU
#asahi-gpu · 2021-05-10 09:52
marcan\o
#asahi-gpu · 2021-05-10 09:52
azonenbergo/
#asahi-gpu · 2021-05-10 08:40
jn__using the address of the parameter key, rather than its pointer value
#asahi-gpu · 2021-05-10 04:08
PthariensFlame[m😅
#asahi-gpu · 2021-05-10 03:12
bloomLet's play spot the bug 🤦‍♀️
#asahi-gpu · 2021-05-10 03:12
bloom...
#asahi-gpu · 2021-05-10 03:12
bloom struct hash_entry *he = _mesa_hash_table_search(so->variants, &key);
#asahi-gpu · 2021-05-10 03:11
bloom....
#asahi-gpu · 2021-05-10 03:11
bloom{
#asahi-gpu · 2021-05-10 03:11
bloomagx_update_shader(struct agx_context *ctx, struct agx_compiled_shader **out, enum pipe_shader_type stage, struct asahi_shader_key *key)
#asahi-gpu · 2021-05-08 16:56
bloomAt this pass just interpret as a fail. But flakes make continuous integration a mess.
#asahi-gpu · 2021-05-08 16:55
bloomjix: Running the test multiple times in a row and sometimes pass, sometimes fail
#asahi-gpu · 2021-05-08 16:55
jixwhat does flake mean here?

plus ›