wip

Python/taichipi.py

+#
+# taichipi.py
+# Neil Gershenfeld 2/22/23
+# calculation of pi by Taichi
+# pi = 3.14159265358979323846 
+#
+import time
+import taichi as ti
+
+NPTS = 10000000
+
+def calc_pi(NPTS):
+   pi = 0
+   for i in range(1,(NPTS+1)):
+      pi += 0.5/((i-0.75)*(i-0.25))
+   return pi
+
+pi = calc_pi(1000) # run first to compile
+start_time = time.time()
+pi = calc_pi(NPTS)
+end_time = time.time()
+mflops = NPTS*5.0/(1.0e6*(end_time-start_time))
+print("\nNPTS = %d, pi = %f"%(NPTS,pi))
+print("time = %f, estimated Python MFlops = %f\n"%(end_time-start_time,mflops))
+
+ti.init(arch=ti.cpu,default_ip=ti.i64,default_fp=ti.f64,cpu_max_num_threads=1)
+
+NPTS = int(1000000000)
+
+@ti.kernel
+def calc_pi_CPU(NPTS:int)->float:
+   pi = float(0)
+   for i in range(1,(NPTS+1)):
+      pi += 0.5/((i-0.75)*(i-0.25))
+   return pi
+
+pi = calc_pi_CPU(1000) # run first to compile
+start_time = time.time()
+pi = calc_pi_CPU(NPTS)
+end_time = time.time()
+mflops = NPTS*5.0/(1.0e6*(end_time-start_time))
+print("NPTS = %d, pi = %f"%(NPTS,pi))
+print("time = %f, estimated Taichi CPU serial MFlops = %f\n"%(end_time-start_time,mflops))
+
+ti.init(arch=ti.cpu,default_ip=ti.i64,default_fp=ti.f64)
+
+@ti.kernel
+def calc_pi_CPU(NPTS:int)->float:
+   pi = float(0)
+   for i in range(1,(NPTS+1)):
+      pi += 0.5/((i-0.75)*(i-0.25))
+   return pi
+
+pi = calc_pi_CPU(1000) # run first to compile
+start_time = time.time()
+pi = calc_pi_CPU(NPTS)
+end_time = time.time()
+mflops = NPTS*5.0/(1.0e6*(end_time-start_time))
+print("NPTS = %d, pi = %f"%(NPTS,pi))
+print("time = %f, estimated Taichi CPU parallel MFlops = %f\n"%(end_time-start_time,mflops))
+
+ti.init(arch=ti.gpu,default_ip=ti.i64,default_fp=ti.f64)
+
+@ti.kernel
+def calc_pi_GPU(NPTS:int)->float:
+   pi = float(0)
+   for i in range(1,(NPTS+1)):
+      pi += 0.5/((i-0.75)*(i-0.25))
+   return pi
+
+pi = calc_pi_GPU(1000) # run first to compile
+start_time = time.time()
+pi = calc_pi_GPU(NPTS)
+end_time = time.time()
+mflops = NPTS*5.0/(1.0e6*(end_time-start_time))
+print("NPTS = %d, pi = %f"%(NPTS,pi))
+print("time = %f, estimated Taichi GPU MFlops = %f\n"%(end_time-start_time,mflops))

README.md

@@ -13,6 +13,7 @@
 |1,635|[cudapi.cu](CUDA/cudapi.cu)|C++, CUDA, 5120 cores|NVIDIA V100|March, 2020|
 |1,595|prior|IBM Blue Gene/P|C, MPI, 4096 processes|prior|
 |1,090|[numbapig.py](Python/numbapig.py)|Python, Numba, CUDA, 5120 cores|NVIDIA V100|March, 2020|
+|1,062|[taichipi.py](Python/taichipi.py)|Python, Taichi, 5120 cores|NVIDIA V100|March, 2023|
 |811|prior|Cray XT4|C, MPI, 2048 processes|prior|
 |315|[numbapip.py](Python/numbapip.py)|Python, Numba, parallel, fastmath<br>96 cores|Intel 2x Xeon Platinum 8175M|February, 2020|
 |272|[threadpi.c](C/threadpi.c)|C, 96 threads<br>gcc threadpi.c -o threadpi -O3 -ffast-math -pthread|Intel 2x Xeon Platinum 8175M|June, 2019|