Changes between Version 23 and Version 24 of Tests_With_RV870

Timestamp:

Oct 8, 2009 1:12:39 AM (15 years ago)

Author:

nakasato

Comment:

--

Tests_With_RV870

@@ -21 +21 @@
 
 == Analysis of VLIW instructions ==
-Our kernel is directly written in IL. After compilation into a machine code, I analyze the generated VLIW instructions. Results are obtained with fglrx 8.65.4 (Catalyst 9.9) and CAL 1.4beta. Almost 90% of the time, "5870_opt kernel" runs at maximum efficiency (here, I mean more than 4 slots are occupied by some instructions). It is compared to 83% in the case of our old kernel. 
+Our kernel is directly written in IL. After compilation into a machine code, I analyze the generated VLIW instructions. Results are obtained with fglrx 8.65.4 (Catalyst 9.9) and CAL 1.4beta. 
+
+Both RV770 and RV870 archtecture has 5-way VLIW units at its heart. Depnding on detailed computations, it is supposed that the device driver (or internal compiler?) tries to fill 5-slots as much as possible. The first column shows a number of occupied slots and the second column shows a number of corresponding VLIW instructions appeared in the kernel. The third column indicates a fraction of VLIW instructions with a given number of slots. The last row presents a total numberf VLIW instructions in the kernel.
 
 === 5870_opt kernel ===
@@ -41 +43 @@
 ||total|| 81|| ||
 
+Almost 90% of the time, "5870_opt kernel" runs at maximum efficiency (here, I mean more than 4 slots are occupied by some operations). It is compared to 83% in the case of our old kernel. 
+
 == Benchmark system ==
 || CPU || Core2 E8400 3.0 GHz ||