flamenco, vm: implement increase_cpi_account_info_limit

[topointon-jump]

Implementing increase_cpi_account_info_limit (solana-foundation/solana-improvement-documents#339), which improves UX/reduces work for programs which make CPI calls that were themselves invoked with more than 64 accounts - they no longer have to de-dup the list of account infos to make a CPI call.

This SIMD also starts charging for serialized account infos and instruction account metas in CPIs.

• Implementation
• Unit tests
• Ledger tests
• Fuzzing

[github-actions]

Performance Measurements ⏳

Suite	Baseline	New	Change
backtest mainnet-368528500-perf per slot	0.076551 s	0.076857 s	0.400% ✅
backtest mainnet-368528500-perf snapshot load	4.866 s	3.104 s	-36.210% ✅
backtest mainnet-368528500-perf total elapsed	76.551297 s	76.857072 s	0.399% ✅
firedancer mem usage with mainnet.toml	1013.23 GiB	1013.23 GiB	0.000% ✅

[github-actions]

Performance Measurements ⏳

Suite	Baseline	New	Change
backtest mainnet-368528500-perf per slot	0.077268 s	0.076864 s	-0.523% ✅
backtest mainnet-368528500-perf snapshot load	4.876 s	3.09 s	-36.628% ✅
backtest mainnet-368528500-perf total elapsed	77.267706 s	76.864491 s	-0.522% ✅
firedancer mem usage with mainnet.toml	1013.23 GiB	1013.23 GiB	0.000% ✅

[github-actions]

Performance Measurements ⏳

Suite	Baseline	New	Change
backtest mainnet-368528500-perf per slot	0.050437 s	0.051482 s	2.072% ✅
backtest mainnet-368528500-perf snapshot load	1.637 s	1.623 s	-0.855% ✅
backtest mainnet-368528500-perf total elapsed	50.436862 s	51.481564 s	2.071% ✅
firedancer mem usage with mainnet.toml	1005.23 GiB	1005.23 GiB	0.000% ✅

[mjain-jump]

looks good, is it fuzz tested?

[github-actions]

Performance Measurements ⏳

Suite	Baseline	New	Change
backtest mainnet-368528500-perf per slot	0.050809 s	0.050696 s	-0.222% ✅
backtest mainnet-368528500-perf snapshot load	1.635 s	1.64 s	0.306% ✅
backtest mainnet-368528500-perf total elapsed	50.809322 s	50.695608 s	-0.224% ✅
firedancer mem usage with mainnet.toml	1005.23 GiB	1005.23 GiB	0.000% ✅

[github-actions]

Performance Measurements ⏳

Suite	Baseline	New	Change
backtest mainnet-368528500-perf per slot	0.050689 s	0.050499 s	-0.375% ✅
backtest mainnet-368528500-perf snapshot load	1.65 s	1.675 s	1.515% ✅
backtest mainnet-368528500-perf total elapsed	50.689069 s	50.499078 s	-0.375% ✅
firedancer mem usage with mainnet.toml	1005.23 GiB	1005.23 GiB	0.000% ✅

[github-actions]

Performance Measurements ⏳

Suite	Baseline	New	Change
backtest mainnet-368528500-perf per slot	0.050624 s	0.050716 s	0.182% ✅
backtest mainnet-368528500-perf snapshot load	1.638 s	1.641 s	0.183% ✅
backtest mainnet-368528500-perf total elapsed	50.623726 s	50.715631 s	0.182% ✅
firedancer mem usage with mainnet.toml	1005.23 GiB	1005.23 GiB	0.000% ✅

[topointon-jump]

VM CPI fuzzer has been running for a while and not showing any bugs