#1 - Update README.md

README.md

@@ -145,28 +145,33 @@ mp: 0:2610744987 p: 1050 m: 0:966616035460 ms: 67696 dur: 182
 
 ## GPU
 20250116: GPU code is CPU bound for thread processing - at 100% cpu (other process) GPU slows by 4-10 times.  I need to increase the threads sent to the GPU past 64k
+
 ### Multi Threaded : 40 bit run 
 #### 128 bit native
 ##### CUDA 12.6: CPP
--  sec 14900K d RTX-A6000 single 60% GPU 54% TDP .5g/48g - 35840k threads / 256 threads/block batch 20 no av
+- sec 13900K b 32 core RTX-4090 Ada single 16384 cores 50% GPU 24% TDP 35840 threads 256 threads/block no av - batch 20
+- 5347 sec 14900K d RTX-A6000 single 60% GPU 54% TDP .5g/48g - 35840k threads / 256 threads/block batch 20 no av
 
 ### Multi Threaded : 37 bit run 
 #### 128 bit native
 ##### CUDA 12.6: CPP
-- 658 sec 14900K d RTX-A6000 single 55% GPU 45% TDP .5g/48g - 35840k threads / 256 threads/block no av
+- 528 sec 13900K b 32 core RTX-4090 Ada single 16384 cores 50% GPU 24% TDP 35840 threads 256 threads/block no av - batch 20
+- 658 sec 14900K d RTX-A6000 single 55% GPU 45% TDP .5g/48g - 35840k threads / 256 threads/block no av - batch 20?
+- 
 ### Multi Threaded : 32 bit run (search 0-(2^32-1) odd integer space)
 #### 128 bit native
 ##### CUDA 12.6: CPP
 - 14 sec 14900K d RTX-A6000 single 55% GPU 45% TDP .5g/48g - 32k threads / 512 threads/block
-- 14 sec RTX-4090 single 16384 cores 48% GPU 24% TDP- 20 batch 40960 threads 512 threads/block 80 blocks
+- 14 sec RTX-4090 Ada single 16384 cores 48% GPU 24% TDP 40960 threads 512 threads/block 80 blocks - batch 20
+- 17 sec RTX-4090 Ada single 16384 cores 48% GPU 24% TDP 35840 threads 256/512 threads/block 160 blocks - batch 20
 - 17 sec P1Gen6 13800H RTX-3500 Ada mobile 5120 cores 60% GPU - 20 batch, 40960 threads
 - 18 sec RTX-A4500 single
 - 18 sec RTX-A4000 single
 - 20 sec RTX-5000 TU104 16g mobile P17gen1
 #### 64 bit native
 Sec: 4 GlobalMax: 319804831 : 1414236446719942480 last search : 1073741825
 -  9 sec 14900K d RTX-A6000 8/32c single 45% GPU 24% TDP .9g/48g - 32k threads / 256 threads/block
-- 10 sec 13900K b 4090 single 45% GPU 22% TDP .9g/24g 32k threads / 256 threads/block 
+- 10 sec 13900K b RTX-4090 Ada single 45% GPU 22% TDP .9g/24g 32k threads / 256 threads/block 
 - 12 sec 13900K a RTX-A4000 single 45% GPU 58% TDP .9g/16g
 
 ## CPU
@@ -318,7 +323,7 @@ large batch 12-14 up from 5 sizes for larger memory 64-128g, cpu for Pcores goes
 - 689 sec 13900K a 3.0/5.7 GHz
 
 # Records stats
-## 128 bit CUDA (5120 to 32768 cores) - RTX-A6000 or dual RTX-4090
+## 128 bit CUDA (5120 to 32768 cores) - RTX-A6000 or dual RTX-4090 Ada
 55% GPU at 24% TDP
 
 ```