M-Labs
/
zynq-rs
10
0
Fork 12
Code Issues 6 Pull Requests 1 Releases Wiki Activity

l2c related changes #76

Merged
sb10q merged 2 commits from pca006132/zynq-rs:master into master 2021-01-26 12:01:24 +08:00
Conversation 1 Commits 2 Files Changed 3
pca006132 commented 2021-01-26 11:22:54 +08:00
  1. Enabled some options to optimize memcpy throughput. Tests on zeus shows minor improvement in RPC throughput.
  2. Removed L2C enable code from SZL, move to artiq-zynq to stay compatible with FSBL.

Note: Some CPU options for performance optimizations are still missing from FSBL.

1. Enabled some options to optimize `memcpy` throughput. Tests on zeus shows minor improvement in RPC throughput. 2. Removed L2C enable code from SZL, move to artiq-zynq to stay compatible with FSBL. Note: Some CPU options for performance optimizations are still missing from FSBL.
pca006132 added 2 commits 2021-01-26 11:22:54 +08:00
pca006132 commented 2021-01-26 11:24:27 +08:00
Poster

Reference:

By default, the prefetch offset is 5'b00000. For example, if S0 receives a cacheable read at address 0x100, the cache line at address 0x120 is prefetched. Prefetching the following cache line might not result in optimal performance. In some systems, it might be better to prefetch more in advance to achieve better performance. The prefetch offset enables this by setting the address of the prefetched cache line to Cache Line + 1 + Offset. The optimal value of the prefetch offset depends on the L3 read latency and on the L1 read issuing capability. It is recommended to perform system experiments by varying the prefetch offset, to find the optimal value.

https://developer.arm.com/documentation/ddi0246/d/functional-overview/implementation-details/prefetching-operation

Reference: > By default, the prefetch offset is 5'b00000. For example, if S0 receives a cacheable read at address 0x100, the cache line at address 0x120 is prefetched. Prefetching the following cache line might not result in optimal performance. In some systems, it might be better to prefetch more in advance to achieve better performance. The prefetch offset enables this by setting the address of the prefetched cache line to Cache Line + 1 + Offset. The optimal value of the prefetch offset depends on the L3 read latency and on the L1 read issuing capability. It is recommended to perform system experiments by varying the prefetch offset, to find the optimal value. https://developer.arm.com/documentation/ddi0246/d/functional-overview/implementation-details/prefetching-operation
sb10q merged commit 6e6612bc3e into master 2021-01-26 12:01:24 +08:00
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels
No items
No Label
Milestone
Clear milestone
No items
No Milestone
Assignees
Clear assignees
No Assignees
1 Participants
Notifications
Due Date
The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Dependencies

No dependencies set.

Reference: M-Labs/zynq-rs#76
There is no content yet.
Delete Branch "pca006132/zynq-rs:master"

Deleting a branch is permanent. Although the deleted branch may exist for a short time before cleaning up, in most cases it CANNOT be undone. Continue?