@@ -52,51 +52,58 @@ _The following are changes which have been implemented in the VTR master branch
52
52
53
53
### Added
54
54
* Support for Advanced Architectures:
55
- * 3D FPGA architectures.
56
- * Architectures with hard Network -on-Chip (NoC ).
57
- * Configurable horizontal and vertical channel widths and types.
58
- * Diagonal routing wires and other complex wire shapes.
55
+ * 3D FPGA and RAD architectures.
56
+ * Architectures with hard Networks -on-Chip (NoCs ).
57
+ * Distinct horizontal and vertical channel widths and types.
58
+ * Diagonal routing wires and other complex wire shapes (L-shaped, T-shaped, ....) .
59
59
60
60
* New Benchmark Suites:
61
- * Koios: A deep-learning-focused benchmark suite.
61
+ * Koios: A deep-learning-focused benchmark suite with various design sizes .
62
62
* Hermes: Benchmarks utilizing hard NoCs.
63
- * TitanNew: Benchmarks targeting the Stratix 10 architecture.
63
+ * TitanNew: Large benchmarks targeting the Stratix 10 architecture.
64
64
65
- * Enhanced Architecture Capture :
65
+ * Commercial FPGAs Architecture Captures :
66
66
* Intel’s Stratix 10 FPGA architecture.
67
67
* AMD’s 7-series FPGA architecture.
68
68
69
- * Parmys Frontend Flow:
70
- * Better Verilog and SystemVerilog language coverage
69
+ * Parmys Logic Synthesis Flow:
70
+ * Better Verilog language coverage
71
71
* More efficient hard block mapping
72
72
73
73
* VPR Graphics Visualizations:
74
- * New interface for improved usability.
75
- * Breakpoint visualizations for placement and routing algorithm debugging.
74
+ * New interface for improved usability and underlying graphics rewritten using EZGL/GTK to allow more UI widgets .
75
+ * Algorithm breakpoint visualizations for placement and routing algorithm debugging.
76
76
* User-guided (manual) placement optimization features.
77
+ * Enabled a live connection for client graphical application to VTR engines through sockets (server mode).
78
+ * Interactive timing path analysis (IPA) client using server mode.
77
79
78
80
* Performance Enhancements:
79
- * Parallel router for faster inter-cluster routing.
81
+ * Parallel router for faster inter-cluster routing or flat routing .
80
82
81
83
* Re-clustering API to modify packing decisions during the flow.
82
84
* Support for floorplanning and placement constraints.
83
- * Unified intra- and inter-cluster routing.
84
- * Comprehensive web-based VTR utilities and APIs documentation.
85
-
85
+ * Unified intra- and inter-cluster (flat) routing.
86
+ * Comprehensive web-based VTR utilities and API documentation.
87
+
86
88
### Changed
87
- * The default values of many commandline options (e.g. inner_num is 0.5 instead of 1.0)
89
+ * The default values of many command line options (e.g. inner_num is 0.5 instead of 1.0)
88
90
* Changes to placement engine
89
- * Smart centroid initial placement algorithm
90
- * Multiple smart placement directed moves
91
- * Reinforcement learning-based placement algorithm
91
+ * Smart centroid initial placement algorithm.
92
+ * Multiple smart placement directed moves.
93
+ * Reinforcement learning-based placement algorithm.
94
+ * Changes to routing engine
95
+ * Faster lookahead creation.
96
+ * More accurate lookahead for large blocks.
97
+ * More efficient heap and pruning strategies.
98
+ * max ` pres_fac ` capped to avoid possible numeric issues.
99
+
92
100
93
101
### Fixed
94
102
* Many algorithmic and coding bugs are fixed in this release
95
-
96
- ### Deprecated
97
-
103
+
98
104
### Removed
99
-
105
+ * Breadth-first (non-timing-driven) router.
106
+ * Non-linear congestion placement cost.
100
107
101
108
## v8.0.0 - 2020-03-24
102
109
0 commit comments