adds a performance section to the homepage #425
Draft
+169
−71
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
Signed-off-by: Kyle J. Davis <[email protected]>
Signed-off-by: Kyle J. Davis <[email protected]>
Signed-off-by: Roshan Khatri <[email protected]>
Member
|
I have updated the PR with a Followed by the /performance/ page: Please add your feedback, I will also update the iframe links once we are ready to merge |
|
Super cool. Would be great to link to relevant blogs to performance and hash tables in case readers want to dive deeper. |
madolson
reviewed
Dec 11, 2025
_data/perf.toml
Outdated
| title = "Throughput Across Versions" | ||
| iframe_url = "https://df0m5orgq2d38.cloudfront.net/public-dashboards/e6e8ec1f88954ddd8a67df1ab4966eb2" | ||
| description = "This dashboard visualizes throughput trends across Valkey versions. It helps compare key releases side by side, highlight performance gains from new features. Use it when evaluating whether a new version meets your performance SLOs before rollout." | ||
| methodology = "These metrics are generated using the valkey-perf-benchmark tool on a c8g.metal.2xl instance, running a matrix of configurations that vary pipelining (1, 10), I/O threading (1, 9), and data sizes 512 bytes. To further stabilize results, we apply IRQ tuning by pinning network interrupts away from CPUs dedicated to the server process, and we isolate the server and benchmark client on separate NUMA nodes to remove L3 cache contention. This controlled methodology ensures consistent, repeatable comparisons across versions and makes performance gains or regressions easier to attribute to real code changes rather than environmental variability." |
Member
There was a problem hiding this comment.
Suggested change
| methodology = "These metrics are generated using the valkey-perf-benchmark tool on a c8g.metal.2xl instance, running a matrix of configurations that vary pipelining (1, 10), I/O threading (1, 9), and data sizes 512 bytes. To further stabilize results, we apply IRQ tuning by pinning network interrupts away from CPUs dedicated to the server process, and we isolate the server and benchmark client on separate NUMA nodes to remove L3 cache contention. This controlled methodology ensures consistent, repeatable comparisons across versions and makes performance gains or regressions easier to attribute to real code changes rather than environmental variability." | |
| methodology = "These metrics are generated using the valkey-perf-benchmark tool on an AWS c8g.metal.24xl instance, running a matrix of configurations that vary pipelining (1, 10), I/O threading (1, 9), and with a data size of 512 bytes. To further stabilize results, we apply IRQ tuning by pinning network interrupts away from CPUs dedicated to the server process, and we isolate the server and benchmark client on separate NUMA nodes to remove L3 cache contention." |
Include a link to the valkey-perf-benchmark tool.
The last sentence adds nothing.
Member
There was a problem hiding this comment.
We also don't test the full matrix, since IO thread 9 and Pipelining 1 is not present. Is that intentional?
Member
There was a problem hiding this comment.
We do have that data, but thats intentional, I can add it
_data/perf.toml
Outdated
| [[sections]] | ||
| title = "Memory Overhead Across Versions" | ||
| iframe_url = "https://df0m5orgq2d38.cloudfront.net/public-dashboards/1e3d87a62b794ce9aa81281d8399e4bf" | ||
| description = "This dashboard visualizes Memory Efficiency trends across Valkey versions. It helps compare releases side by side, highlight Memory Efficiency gains from new features. Use it when evaluating whether a new version meets your performance SLOs before rollout." |
Member
There was a problem hiding this comment.
Suggested change
| description = "This dashboard visualizes Memory Efficiency trends across Valkey versions. It helps compare releases side by side, highlight Memory Efficiency gains from new features. Use it when evaluating whether a new version meets your performance SLOs before rollout." | |
| description = "This dashboard visualizes Memory Efficiency trends across Valkey versions. It helps compare releases side by side, highlight Memory Efficiency gains from new features." |
What Performance SLO?
Member
There was a problem hiding this comment.
I meant Service Level Objective, but yeah can be removed
| description = "This dashboard visualizes throughput trends across Valkey versions. It helps compare key releases side by side, highlight performance gains from new features. Use it when evaluating whether a new version meets your performance SLOs before rollout." | ||
| methodology = "These metrics are generated using the valkey-perf-benchmark tool on a c8g.metal.2xl instance, running a matrix of configurations that vary pipelining (1, 10), I/O threading (1, 9), and data sizes 512 bytes. To further stabilize results, we apply IRQ tuning by pinning network interrupts away from CPUs dedicated to the server process, and we isolate the server and benchmark client on separate NUMA nodes to remove L3 cache contention. This controlled methodology ensures consistent, repeatable comparisons across versions and makes performance gains or regressions easier to attribute to real code changes rather than environmental variability." | ||
|
|
||
| [[sections]] |
Member
There was a problem hiding this comment.
Memory overhead is not performance. I'm not sure why we're including this here.
_data/perf.toml
Outdated
| title = "Memory Overhead Across Versions" | ||
| iframe_url = "https://df0m5orgq2d38.cloudfront.net/public-dashboards/1e3d87a62b794ce9aa81281d8399e4bf" | ||
| description = "This dashboard visualizes Memory Efficiency trends across Valkey versions. It helps compare releases side by side, highlight Memory Efficiency gains from new features. Use it when evaluating whether a new version meets your performance SLOs before rollout." | ||
| methodology = "These benchmarks are generated by starting an empty instance of Valkey for each test, measuring memory usage, then adding 3 million string items of a certain data size, then measuring Valkey memory usage again. We take the increase in memory use, divide by the number of items, and subtract the size of the user data (key and value) to get the extra overhead bytes Valkey uses to track and organize the data. For this graph, we tested every value size from 8B to 128B inclusive, then averaged the numbers for each range in this chart." |
Member
There was a problem hiding this comment.
I'm still confused why this isn't an integration test. Why is this here? What is this methodology?
Member
There was a problem hiding this comment.
This was with the intention of showing how we have improved valkey for throughput and memory utilisation
Member
There was a problem hiding this comment.
We can surely add an integration test for it too
_data/perf.toml
Outdated
| [[sections]] | ||
| title = "Throughput Across Versions" | ||
| iframe_url = "https://df0m5orgq2d38.cloudfront.net/public-dashboards/e6e8ec1f88954ddd8a67df1ab4966eb2" | ||
| description = "This dashboard visualizes throughput trends across Valkey versions. It helps compare key releases side by side, highlight performance gains from new features. Use it when evaluating whether a new version meets your performance SLOs before rollout." |
Member
There was a problem hiding this comment.
Suggested change
| description = "This dashboard visualizes throughput trends across Valkey versions. It helps compare key releases side by side, highlight performance gains from new features. Use it when evaluating whether a new version meets your performance SLOs before rollout." | |
| description = "This dashboard visualizes throughput trends across Valkey versions. It helps compare key releases side by side, highlight performance gains from new features." |
Performance SLOs is a very weird thing to call out.
Signed-off-by: Roshan Khatri <[email protected]>
Member
|
Applied the suggestions and also added the links to the relevant blogs. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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.
Description
Issues Resolved
#421
Check List
--signoffBy submitting this pull request, I confirm that my contribution is made under the terms of the BSD-3-Clause License.