[RouterLookahead] Router Lookahead Report #3161
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Here is an example of what the report looks like: 
... 
|
|
@amin1377 FYI, you might find this option interesting. |
Added an option to generate a report on the router lookahead. This reports profiles the router lookahead by performing a set of trial routes from source RR nodes to all target RR nodes and comparing the estimated cost from the router lookahead to the actual cost of the path. Notably, this report tries to show the mean squared error of the router lookahead and the maximum overestimation of the router lookahead. The mean squared error is a measure of how accurate the router lookahead is. The more accurate the router lookahead is, the faster the path search will be (with less loss in quality). The max overestimation is a measure of the admissibility of the router lookahead heuristic. The more the router lookahead overestimates, the worse the quality of the path found (theoretically).
AlexandreSinger
force-pushed
the
feature-lookahead-profile-report
branch
from
d4026a5 to
c60fde5
Compare
github-actions
bot
added
VPR
| return "UNKNOWN"; | ||
| } | ||
| } | ||
|
|
There was a problem hiding this comment.
rr_node_typename in rr_node_types.h does the same thing.
|
|
||
| // Create a NOP router lookahead to be used during the all-destination dijkstra search. | ||
| t_det_routing_arch temp_det_routing_arch; | ||
| auto temp_router_lookahead = make_router_lookahead(temp_det_routing_arch, e_router_lookahead::NO_OP, |
There was a problem hiding this comment.
replace auto with explicit type name
| float heuristic_delay = 0.0f; | ||
| float path_delay = 0.0f; | ||
| }; | ||
| std::vector<t_overestimation_info> max_overestimation_per_type((int)e_rr_type::NUM_RR_TYPES); |
There was a problem hiding this comment.
You can use vtr::vector and index it directly with e_rr_type
| * valuable since over-estimations in the router lookahead cause the | ||
| * heuristic to be inadmissible which can hurt quality. | ||
| */ | ||
| static void profile_lookahead_overestimation(std::ofstream& os, |
There was a problem hiding this comment.
This functio is quite long. Consider breaking it down
| // The target number of random source (sample) nodes to use. | ||
| size_t target_num_trials = 100; | ||
| // The maximum number of attempts. | ||
| size_t max_attempts = 10000; |
There was a problem hiding this comment.
use constexpr if there is no plan to modify these values through command line options
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Added an option to generate a report on the router lookahead. This reports profiles the router lookahead by performing a set of trial routes from source RR nodes to all target RR nodes and comparing the estimated cost from the router lookahead to the actual cost of the path.
Notably, this report tries to show the mean squared error of the router lookahead and the maximum overestimation of the router lookahead.
The mean squared error is a measure of how accurate the router lookahead is. The more accurate the router lookahead is, the faster the path search will be (with less loss in quality). The max overestimation is a measure of the admissibility of the router lookahead heuristic. The more the router lookahead overestimates, the worse the quality of the path found (theoretically).