sweet: remove biogo benchmarks from default set

These benchmarks have shown themselves to not be of a lot of use over
time. They tend to exhibit pathological behaviors not visible in other
programs, that could be worked around if the benchmarks themselves were
improved. Note that these benchmarks are, in part, written by us, using
the biogo library.

biogo-igor is also causing flaky test failures, so this CL skips them for
the end-to-end test, though we think it might be a real PGO bug.

Let's drop them at this point. We'll add back something more interesting
very soon.

Fixes golang/go#56958.
Updates golang/go#69614.

Change-Id: Ic3761953abf4c8db19b3623e135a809c841cf160
Reviewed-on: https://go-review.googlesource.com/c/benchmarks/+/615416
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Michael Pratt <mpratt@google.com>

sweet/cmd/sweet/benchmark.go

@@ -98,8 +98,6 @@ var benchmarkGroups = func() map[string][]*benchmark {
 	m := make(map[string][]*benchmark)
 
 	m["default"] = []*benchmark{
-		allBenchmarksMap["biogo-igor"],
-		allBenchmarksMap["biogo-krishna"],
 		allBenchmarksMap["bleve-index"],
 		allBenchmarksMap["cockroachdb"],
 		allBenchmarksMap["etcd"],

sweet/cmd/sweet/integration_test.go

@@ -207,8 +207,6 @@ func TestSweetEndToEnd(t *testing.T) {
 	for i, shard := range []shard{
 		{"tile38", 2},
 		{"go-build", 4},
-		{"biogo-igor", 1},
-		{"biogo-krishna", 1},
 		{"cockroachdb", 1},
 		{"etcd", 1},
 		{"bleve-index", 1},