Discover and read the best of Twitter Threads about #LLVM
Most recents (4)
@boblautenbach @Alan_L_Lovejoy @Ryu_Zoku Yes Chris Lattner is very interesting, been following his #LLVM and #SWIFT work.
1/
1/
@boblautenbach @Alan_L_Lovejoy @Ryu_Zoku Actually #ZokuTalk adopts the portion of SWIFT #EBNF for #Unicode as they seems have done good job on that, as well as parts of the EBNF for numerics such as "1_234.00" to expand privative data types entry.
\2
\2
@boblautenbach @Alan_L_Lovejoy @Ryu_Zoku Borrowing C & Assembly data types: 0b0011, 0xFF, keeping 16r0A, plus many others.
Actually the numeric formats are one of more complicated part of the #EBNF of #ZokuTalk and the only non-#MOBS (#Messages #Objects #Block #Syntax) part of the evolved syntax.
\3
Actually the numeric formats are one of more complicated part of the #EBNF of #ZokuTalk and the only non-#MOBS (#Messages #Objects #Block #Syntax) part of the evolved syntax.
\3
#CLANG CAN BE BUILT WITHIN ONE MINUTE #LLVM
[2564/2564] Creating executable symlink bin/clang
real 0m56.685s
user 75m20.548s
* AWS c5a.24xlarge (96 vcpus)
* target=x86 MinSizeRel
* Using PGO+LTO clang and lld (up to 30% faster)
The trace.json is here; gist.github.com/chapuni/6b80af…
[2564/2564] Creating executable symlink bin/clang
real 0m56.685s
user 75m20.548s
* AWS c5a.24xlarge (96 vcpus)
* target=x86 MinSizeRel
* Using PGO+LTO clang and lld (up to 30% faster)
The trace.json is here; gist.github.com/chapuni/6b80af…
Tweaks in CMake part of the LLVM tree
* Enhanced LLVM_ENABLE_MODULES
** Generator and targets for explicit modules
* Modularized tblgen (invokes particular tblgen-emitters.so)
* Introduce obj.LLVMSupportLite for tblgen
* KILL target_link_libraries(PUBLIC) !!!IMPORTANT!!!
* Enhanced LLVM_ENABLE_MODULES
** Generator and targets for explicit modules
* Modularized tblgen (invokes particular tblgen-emitters.so)
* Introduce obj.LLVMSupportLite for tblgen
* KILL target_link_libraries(PUBLIC) !!!IMPORTANT!!!
Tweaks to clang
* Introduce an option "Include headers textually instead of waiting for building modules"
* Introduce an option "Include headers textually instead of waiting for building modules"
C++ vtables are heavily used to dispatch virtual function calls. Attackers routinely hijack vtables to divert control flow. In our recent @IEEEEUROSP '21 paper by @MarkusBauer_mkb, we show that statically-linked programs can live (much safer) *without* such vtables. (1/3)
We radically change vtable defenses. We release #NoVT, an #LLVM-based compiler that *replaces* vtables by switch-case constructs that are inherently control-flow safe. To this end, each C++ class gets unique IDs which are used to dispatch the correct method implementations. (2/3)
Instead of protecting vtables, NoVT eliminates them. This not only is simpler, but also more efficient. NoVT is the first vtable defense that does not degrade performance (on average)! Code github.com/novt-vtable-le… and paper publications.cispa.saarland/3364/1/NoVT-fo… here. (3/3)
I'm about to buy a new MacBook Pro 13" with the famous Apple M1 SoC.
Let's see how ready it is for developers.
A thread ⤵️
Legend: 🔴 = not working, 🟢 = working
#AppleSiliconM1
Let's see how ready it is for developers.
A thread ⤵️
Legend: 🔴 = not working, 🟢 = working
#AppleSiliconM1
🟢 Rust is working!
`aarch64-apple-darwin` is part of the Tier 2 platforms, doc.rust-lang.org/nightly/rustc/….
See the PR, github.com/rust-lang/rust….
#AppleSiliconM1 #rustlang
`aarch64-apple-darwin` is part of the Tier 2 platforms, doc.rust-lang.org/nightly/rustc/….
See the PR, github.com/rust-lang/rust….
#AppleSiliconM1 #rustlang
