add annotations week of 1/23/25 (#26594)

- qthreads partial reversion (impacts thread ring shootout benchmark)
- Add annotation indicating we suspect a PR raising the number of gasnet
threads may have influenced nbody perf on Apollo. If not due to that, it
could also be due to some bimodal machine behavior, graph history makes
it look like such behavior may exist.

[Reviewed by nobody; annotations update]

test/ANNOTATIONS.yaml

@@ -609,6 +609,10 @@ all:
   12/12/24:
     - text: Backport Fix For Performance Regression in Qthreads 1.21 (#26328)
       config: chapcs, 16-node-apollo-hdr, 16-node-hpe-cray-ex, 1-node-hpe-cray-ex
+  01/06/25:
+    - text: Partially revert Qthreads patch from #26328 (#26468)
+      config: chapcs, 16-node-apollo-hdr, 16-node-hpe-cray-ex, 1-node-hpe-cray-ex
+
 
 # End all
 
@@ -1662,6 +1666,9 @@ miniMD.ml-time:
     - Manually hoist loop-invariant array access in miniMD (#7432)
   09/26/17:
     - Prevent certain data races upon 'in' forall intents (#7467)
+  01/06/25:
+    - text: Suspect - Raise number of threads used by GASNet by default + add warning if exceeded (#26448)
+      config: 16-node-apollo-hdr
 
 ml-memleaksfull:
   04/04/20:
@@ -1699,7 +1706,7 @@ nbody:
     - Change param-based nbody to foreach-based (#24745)
   10/06/24:
     - Update my study version of nbody to be like nbody#2, but using foreach loops (#26044)
-
+ 
 no-op:
   02/24/17:
     - upgrade hwloc to 1.11.6 (#5411)