add sdata to control individual sims
This commit is contained in:
Stefan Holst committed
1 parent
87d93afb44
commit
b981b1153c
1 file changed
+44
-20
+44
-20
@@ -122,6 +122,9 @@ class WaveSim:
|
||||
self.lst_eat_valid = False
|
||||
|
||||
self.cdata = np.zeros((len(self.interface), sims, 7), dtype='float32')
|
||||
|
||||
self.sdata = np.zeros((sims, 4), dtype='float32')
|
||||
self.sdata[...,0] = 1.0
|
||||
|
||||
if isinstance(wavecaps, int):
|
||||
wavecaps = [wavecaps] * len(circuit.lines)
|
||||
@@ -328,7 +331,7 @@ class WaveSim:
|
||||
sims = min(sims or self.sims, self.sims)
|
||||
for op_start, op_stop in zip(self.level_starts, self.level_stops):
|
||||
self.overflows += level_eval(self.ops, op_start, op_stop, self.state, self.sat, 0, sims,
|
||||
self.timing, sd, seed)
|
||||
self.timing, self.sdata, sd, seed)
|
||||
self.lst_eat_valid = False
|
||||
|
||||
def wave(self, line, vector):
|
||||
@@ -521,12 +524,12 @@ class WaveSim:
|
||||
|
||||
|
||||
@numba.njit
|
||||
def level_eval(ops, op_start, op_stop, state, sat, st_start, st_stop, line_times, sd, seed):
|
||||
def level_eval(ops, op_start, op_stop, state, sat, st_start, st_stop, line_times, sdata, sd, seed):
|
||||
overflows = 0
|
||||
for op_idx in range(op_start, op_stop):
|
||||
op = ops[op_idx]
|
||||
for st_idx in range(st_start, st_stop):
|
||||
overflows += wave_eval(op, state, sat, st_idx, line_times, sd, seed)
|
||||
overflows += wave_eval(op, state, sat, st_idx, line_times, sdata[st_idx], sd, seed)
|
||||
return overflows
|
||||
|
||||
|
||||
@@ -547,7 +550,7 @@ def rand_gauss(seed, sd):
|
||||
|
||||
|
||||
@numba.njit
|
||||
def wave_eval(op, state, sat, st_idx, line_times, sd=0.0, seed=0):
|
||||
def wave_eval(op, state, sat, st_idx, line_times, sdata, sd=0.0, seed=0):
|
||||
lut, z_idx, a_idx, b_idx = op
|
||||
overflows = int(0)
|
||||
|
||||
@@ -563,9 +566,11 @@ def wave_eval(op, state, sat, st_idx, line_times, sd=0.0, seed=0):
|
||||
if z_cur == 1:
|
||||
state[z_mem, st_idx] = TMIN
|
||||
|
||||
a = state[a_mem, st_idx] + line_times[a_idx, 0, z_cur] * rand_gauss(_seed ^ a_mem ^ z_cur, sd)
|
||||
b = state[b_mem, st_idx] + line_times[b_idx, 0, z_cur] * rand_gauss(_seed ^ b_mem ^ z_cur, sd)
|
||||
|
||||
a = state[a_mem, st_idx] + line_times[a_idx, 0, z_cur] * rand_gauss(_seed ^ a_mem ^ z_cur, sd) * sdata[0]
|
||||
if int(sdata[1]) == a_idx: a += sdata[2+z_cur]
|
||||
b = state[b_mem, st_idx] + line_times[b_idx, 0, z_cur] * rand_gauss(_seed ^ b_mem ^ z_cur, sd) * sdata[0]
|
||||
if int(sdata[1]) == b_idx: b += sdata[2+z_cur]
|
||||
|
||||
previous_t = TMIN
|
||||
|
||||
current_t = min(a, b)
|
||||
@@ -576,15 +581,21 @@ def wave_eval(op, state, sat, st_idx, line_times, sd=0.0, seed=0):
|
||||
if b < a:
|
||||
b_cur += 1
|
||||
b = state[b_mem + b_cur, st_idx]
|
||||
b += line_times[b_idx, 0, z_val ^ 1] * rand_gauss(_seed ^ b_mem ^ z_val ^ 1, sd)
|
||||
thresh = line_times[b_idx, 1, z_val] * rand_gauss(_seed ^ b_mem ^ z_val, sd)
|
||||
b += line_times[b_idx, 0, z_val ^ 1] * rand_gauss(_seed ^ b_mem ^ z_val ^ 1, sd) * sdata[0]
|
||||
thresh = line_times[b_idx, 1, z_val] * rand_gauss(_seed ^ b_mem ^ z_val, sd) * sdata[0]
|
||||
if int(sdata[1]) == b_idx:
|
||||
b += sdata[2+(z_val^1)]
|
||||
thresh += sdata[2+z_val]
|
||||
inputs ^= 2
|
||||
next_t = b
|
||||
else:
|
||||
a_cur += 1
|
||||
a = state[a_mem + a_cur, st_idx]
|
||||
a += line_times[a_idx, 0, z_val ^ 1] * rand_gauss(_seed ^ a_mem ^ z_val ^ 1, sd)
|
||||
thresh = line_times[a_idx, 1, z_val] * rand_gauss(_seed ^ a_mem ^ z_val, sd)
|
||||
a += line_times[a_idx, 0, z_val ^ 1] * rand_gauss(_seed ^ a_mem ^ z_val ^ 1, sd) * sdata[0]
|
||||
thresh = line_times[a_idx, 1, z_val] * rand_gauss(_seed ^ a_mem ^ z_val, sd) * sdata[0]
|
||||
if int(sdata[1]) == a_idx:
|
||||
a += sdata[2+(z_val^1)]
|
||||
thresh += sdata[2+z_val]
|
||||
inputs ^= 1
|
||||
next_t = a
|
||||
|
||||
@@ -636,6 +647,7 @@ class WaveSimCuda(WaveSim):
|
||||
self.d_timing = cuda.to_device(self.timing)
|
||||
self.d_tdata = cuda.to_device(self.tdata)
|
||||
self.d_cdata = cuda.to_device(self.cdata)
|
||||
self.d_sdata = cuda.to_device(self.sdata)
|
||||
|
||||
self._block_dim = (32, 16)
|
||||
|
||||
@@ -650,6 +662,9 @@ class WaveSimCuda(WaveSim):
|
||||
|
||||
def set_line_delay(self, line, polarity, delay):
|
||||
self.d_timing[line, 0, polarity] = delay
|
||||
|
||||
def sdata_to_device(self):
|
||||
cuda.to_device(self.sdata, to=self.d_sdata)
|
||||
|
||||
def assign(self, vectors, time=0.0, offset=0):
|
||||
assert (offset % 8) == 0
|
||||
@@ -676,7 +691,7 @@ class WaveSimCuda(WaveSim):
|
||||
for op_start, op_stop in zip(self.level_starts, self.level_stops):
|
||||
grid_dim = self._grid_dim(sims, op_stop - op_start)
|
||||
wave_kernel[grid_dim, self._block_dim](self.d_ops, op_start, op_stop, self.d_state, self.sat, int(0),
|
||||
sims, self.d_timing, sd, seed)
|
||||
sims, self.d_timing, self.d_sdata, sd, seed)
|
||||
cuda.synchronize()
|
||||
self.lst_eat_valid = False
|
||||
|
||||
@@ -858,7 +873,7 @@ def rand_gauss_dev(seed, sd):
|
||||
|
||||
|
||||
@cuda.jit()
|
||||
def wave_kernel(ops, op_start, op_stop, state, sat, st_start, st_stop, line_times, sd, seed):
|
||||
def wave_kernel(ops, op_start, op_stop, state, sat, st_start, st_stop, line_times, sdata, sd, seed):
|
||||
x, y = cuda.grid(2)
|
||||
st_idx = st_start + x
|
||||
op_idx = op_start + y
|
||||
@@ -869,6 +884,7 @@ def wave_kernel(ops, op_start, op_stop, state, sat, st_start, st_stop, line_time
|
||||
a_idx = ops[op_idx, 2]
|
||||
b_idx = ops[op_idx, 3]
|
||||
overflows = int(0)
|
||||
sdata = sdata[st_idx]
|
||||
|
||||
_seed = (seed << 4) + (z_idx << 20) + (st_idx << 1)
|
||||
|
||||
@@ -882,9 +898,11 @@ def wave_kernel(ops, op_start, op_stop, state, sat, st_start, st_stop, line_time
|
||||
if z_cur == 1:
|
||||
state[z_mem, st_idx] = TMIN
|
||||
|
||||
a = state[a_mem, st_idx] + line_times[a_idx, 0, z_cur] * rand_gauss_dev(_seed ^ a_mem ^ z_cur, sd)
|
||||
b = state[b_mem, st_idx] + line_times[b_idx, 0, z_cur] * rand_gauss_dev(_seed ^ b_mem ^ z_cur, sd)
|
||||
|
||||
a = state[a_mem, st_idx] + line_times[a_idx, 0, z_cur] * rand_gauss_dev(_seed ^ a_mem ^ z_cur, sd) * sdata[0]
|
||||
if int(sdata[1]) == a_idx: a += sdata[2+z_cur]
|
||||
b = state[b_mem, st_idx] + line_times[b_idx, 0, z_cur] * rand_gauss_dev(_seed ^ b_mem ^ z_cur, sd) * sdata[0]
|
||||
if int(sdata[1]) == b_idx: b += sdata[2+z_cur]
|
||||
|
||||
previous_t = TMIN
|
||||
|
||||
current_t = min(a, b)
|
||||
@@ -895,15 +913,21 @@ def wave_kernel(ops, op_start, op_stop, state, sat, st_start, st_stop, line_time
|
||||
if b < a:
|
||||
b_cur += 1
|
||||
b = state[b_mem + b_cur, st_idx]
|
||||
b += line_times[b_idx, 0, z_val ^ 1] * rand_gauss_dev(_seed ^ b_mem ^ z_val ^ 1, sd)
|
||||
thresh = line_times[b_idx, 1, z_val] * rand_gauss_dev(_seed ^ b_mem ^ z_val, sd)
|
||||
b += line_times[b_idx, 0, z_val ^ 1] * rand_gauss_dev(_seed ^ b_mem ^ z_val ^ 1, sd) * sdata[0]
|
||||
thresh = line_times[b_idx, 1, z_val] * rand_gauss_dev(_seed ^ b_mem ^ z_val, sd) * sdata[0]
|
||||
if int(sdata[1]) == b_idx:
|
||||
b += sdata[2+(z_val^1)]
|
||||
thresh += sdata[2+z_val]
|
||||
inputs ^= 2
|
||||
next_t = b
|
||||
else:
|
||||
a_cur += 1
|
||||
a = state[a_mem + a_cur, st_idx]
|
||||
a += line_times[a_idx, 0, z_val ^ 1] * rand_gauss_dev(_seed ^ a_mem ^ z_val ^ 1, sd)
|
||||
thresh = line_times[a_idx, 1, z_val] * rand_gauss_dev(_seed ^ a_mem ^ z_val, sd)
|
||||
a += line_times[a_idx, 0, z_val ^ 1] * rand_gauss_dev(_seed ^ a_mem ^ z_val ^ 1, sd) * sdata[0]
|
||||
thresh = line_times[a_idx, 1, z_val] * rand_gauss_dev(_seed ^ a_mem ^ z_val, sd) * sdata[0]
|
||||
if int(sdata[1]) == a_idx:
|
||||
a += sdata[2+(z_val^1)]
|
||||
thresh += sdata[2+z_val]
|
||||
inputs ^= 1
|
||||
next_t = a
|
||||
|
||||
|
||||
Reference in new issue
Block a user