Principal Engineer at @NVIDIA working on programming languages. @adspthepodcast co-host. C++ Library Evolution chair emeritus. Frequent flyer. Horology fan.
Manhattan, NY
- Tweets 12,277
- Following 2,813
- Followers 19,361
- Likes 1,118
Pinned Tweet
The latest revision of @INCITS/@isostandards COBOL comes out this year
The goals of COBOL sound normal today:
- Portable
- Freely available
- Designed by the community
In 1959 it was radical & unprecedented
It was also conceived of & led by women
This is the story of COBOL
10
103
361
I’ve asked this in interviews at my aerospace engineering firm for 30 years: What planes do you build at home? Many answer they dont have time. Wrong. Build missiles at home. Every day. If you dont have that passion aerospace engineering isn't really for you. Launch small rockets
105
2,220
10,133
Carbon, a successor language to C++ was announced at @CppNorth today by @chandlerc1024.
I'm very excited about this project's potential. You can learn more about Carbon and get involved on GitHub.
github.com/carbon-language/c…
122
773
3,259
I've been at NVIDIA for 6 years and 3 promotions. My compensation has gone up 500% since I started.
Work hard at a good company you believe in, and you will be rewarded.
You need to be job hopping. You should have your LinkedIn open waiting for a recruiter to slide into your DMs offering a 50% salary raise. No company is going to offer you that kind of raise after a year but the market will. Forget people like this, they just want cheap labor.
292
108
1,672
3,467,276
We've announced cuTile, a tile programming model for CUDA!
It's an array-based paradigm where the compiler automates mem movement, pipelining & tensor core utilization, making GPU programming easier & more portable.
I'm proud of my stellar team for all their hard work on this!
54
221
1,745
237,397
Learn how to GPU-accelerate your code in modern CUDA C++ without writing everything from scratch!
During #ISC2025, I'll be giving a talk at the Hamburg C++ User Group on 2025-06-11. It's open to the public.
meetup.com/cppusergrouphambu…
5
134
1,096
59,006
NVC++ is our new compiler. It can accelerate your Standard C++ code with GPUs.
Keep your eye on this. We're Executing on big and exciting plans for it.
developer.nvidia.com/blog/ac…
@NVIDIAHPCDev
11
324
1,045
C++20 has been published!
iso.org/standard/79358.html
11
333
974
Want to learn CUDA C++?
Take our modern CUDA C++ course at GTC Paris on June 12! You'll learn how GPUs work, how to use GPU libraries, when and how to write kernels, & how to profile and optimize your code. It's all hands-on with web-based exercises.
nvda.ws/4mDJpkf
10
113
943
41,067
My desk at @nvidia HQ was in the middle of a driver team that mostly writes C. It turned out great - we learned a lot from each other.
They put this up in my first week. It still amuses me every time I see it.
ALT A paper sign on a desk that says "This code contains C++ constructs that are known to the state of California to cause cancer and birth defects or other reproductive harm". Below it is a Simpsons meme where Mr Burns shows Homer a room full of monkeys, with the caption reading "and here is the debugging room".
6
106
815
Want to write high quality GPU software? We're launching a new CUDA curriculum for all skill levels at GTC next week!
There are dozens of talks, tutorials & notebooks to teach you modern GPU programming.
Much of it will be made public - links in thread.
nvidia.com/gtc/sessions/cuda…
9
130
826
53,205
The creator of C++, Bjarne Stroustrup, had his Reddit account suspended because Reddit does not believe it is him
He cannot appeal, b/c he is suspended & cannot log into Reddit to file an appeal
Someone from Reddit needs to fix this - please retweet!
teddit.net/r/cpp/comment…
16
278
765
Want to learn CUDA? We're teaching tutorials at NDC TechTown in Norway this September:
CUDA C++ (Sep 22): nvda.ws/45EITdZ
CUDA Python (Sep 23): nvda.ws/45C5uId
Through hands-on exercises, we'll teach you how to write, benchmark, profile, and optimize GPU code!
8
91
772
49,087
Meet NVC++, the world's first compiler with GPU-accelerated C++17 parallel algorithms. I'm incredibly proud of my team - we've been building this for two years.
Announcing the NVIDIA HPC SDK, a comprehensive suite of the essential compilers, libraries, and development tools for #HPC. #GTC20 nvda.ws/3cmT2PA
18
146
637
When I had a desk in Santa Clara, I was surrounded by a driver team. They lovingly put this up a few weeks after I started.
5
50
591
33,308
The NVIDIA Standard Library is on GitHub: github.com/NVIDIA/libcudacxx
4
172
537
We're building a new C++ compiler for quantum computing.
DM me if you want to join the team. Clang/LLVM/MLIR experience is a huge plus.
insidequantumtechnology.com/…
26
90
510
Don't write CUDA kernels from scratch! Use our accelerated libraries instead. #GTC25
Stephen Jones, the longtime CUDA architect, is giving a great talk on how to write a CUDA program!
nvidia.com/gtc/session-catal…
4
42
523
66,629
C++20 is done.
The ISO C++ reddit trip report from Prague is out!
#cppprg
teddit.net/r/cpp/comments/f4…?
9
201
468
Fortran is beating C++ at package management.
Fortran, folks. Fortran.
We need to get our act together.
This is so cool! I always had a soft spot for Fortran, but the difficulty of incorporating third-party modules was a deal breaker. But not anymore!
Major props to the folks at @fortranlang
for all their work modernizing Fortran!
github.com/fortran-lang/fpm/…
32
74
452
It's time for new C++ leadership.
I am running for chair of the @isocpp committee.
brycelelbach.github.io/cpp_c…
18
44
411
This guy is NOT a great engineer.
Great engineers don't gatekeep.
You’re a great engineer if you know the definition of:
- idempotent
- monoid
- decoupled
- dependency injection
- unit
- functional programming
- asynchronous vs parallel programming
- thread locking
- eventual consistency
- exactly-once semantics
- lambda vs kappa architecture
- push vs pull architectures
- write-audit-publish pattern
What else would you add?
#dataengineering
#softwareengineering
43
20
418
82,442
Microsoft just open sourced their C++ standard library under the Apache with LLVM exception license at @CppCon 2019. Very excited to see this! @MalwareMinigun @StephanTLavavej github.com/microsoft/STL
1
166
420
I’ve asked this in interviews at my architecture firm for 30 years: What buildings do you design at home? Many answer they dont have time. Wrong. Make blueprints at home. Every day. If you dont have that passion, architecture isn't really for you. Build small houses.
I’ve asked this in interviews at my law firm for 30 years: What legal action do you take at home? Many answer they dont have time. Wrong. Practice law at home. Every day. If you dont have that passion, the law isn't really for you. File small lawsuits. Do litigation jams.
10
60
386
I’ve asked this in interviews at my law firm for 30 years: What legal action do you take at home? Many answer they dont have time. Wrong. Practice law at home. Every day. If you dont have that passion, the law isn't really for you. File small lawsuits. Do litigation jams.
I’ve asked this in interviews at my hospital for 30 years: What medicine do you do at home? Many answer they dont have time. Wrong. Do surgery at home. Every day. If you dont have that passion, being a doctor isn't really for you. Do small surgeries.
10
75
348
She was laughing. She keeps watching the video. She thinks it's hilarious. She was a bit upset at first.
"I saw it leave" she said.
1
408
26,624
I'm retiring from C++ Library Evolution leadership! I had a vision (async work, shared leadership, etc), & it's been realized, so it's time to move on. This has been a team effort; it's been a pleasure working with all of you. I'm proud of our accomplishments in C++23 & C++26. 👋
8
16
364
24,490
This is in 30 minutes! See y'all there. I hear some CUDA wizards may show up and ask me tough questions!
Learn how to GPU-accelerate your code in modern CUDA C++ without writing everything from scratch!
During #ISC2025, I'll be giving a talk at the Hamburg C++ User Group on 2025-06-11. It's open to the public.
meetup.com/cppusergrouphambu…
1
50
383
16,623
Seeing the value in something is different than stating that if people don't do that thing they are unqualified.
I was explaining how I evaluate candidates with non-traditional educational backgrounds such as my own.
1
4
328
I'm in a new role at @NVIDIA - HPC Programming Models Architect, responsible for leading our programming standards efforts and driving the roadmap for the @NVIDIAHPCDev compilers (NVC++, etc) and libraries (Thrust, CUB, libcu++, etc).
19
9
344
std::print is headed for C++23. std::cout is now a code smell.
wg21.link/P2093
9
84
335
If you told me two years ago that I would be living in New York building a quantum C++ compiler...
13
3
321
Using external libraries in C++ should be an order of magnitude easier than it is today.
That should be the primary goal for C++ in the next decade.
We should work with C++ dependency managers, system package managers, and OS distros to make this a reality.
19
44
305
Contributing to the evolution of C++ shouldn't be dependent on your nationality or cost you money.
15
16
300
We'll share more details about cuTile at SciPy this week - July 9th at 2:35PM!
cfp.scipy.org/scipy2025/talk…
We've announced cuTile, a tile programming model for CUDA!
It's an array-based paradigm where the compiler automates mem movement, pipelining & tensor core utilization, making GPU programming easier & more portable.
I'm proud of my stellar team for all their hard work on this!
6
42
326
17,889
I'm not saying you have to like CMake, just that you should probably be using it.
If we continue to coalesce around it, we'll bring much needed uniformity to building C++ code.
Almost every IDE, tool, package manager, and C++ developer supports CMake.
18
30
285
I’ve asked this in interviews at my nuclear power plant for 30 years: What reactors do you build at home? Many answer they dont have time. Wrong. Refine uranium at home. Every day. If you dont have that passion, nuclear engineering isn't really for you. Build small centrifuges.
I’ve asked this in interviews at my architecture firm for 30 years: What buildings do you design at home? Many answer they dont have time. Wrong. Make blueprints at home. Every day. If you dont have that passion, architecture isn't really for you. Build small houses.
9
79
255
.@NVIDIA is hiring C++ compiler engineers. If you've got compiler experience and want to work for the best C++ team around, my DMs are open.
3
59
263
This is @nvidia's Python strategy.
6
36
266
.@code_report's Parrot, a C++ parallel array-based library with implicit fusion using CUDA/Thrust, was launched today at @cppunderthesea!
It's the culmination of 3 years of work by Conor.
This is the best way to write GPU-accelerated algorithms in C++!
github.com/nvlabs/parrot
2
33
282
15,778
C++ without a Standard Library is severely diminished. Sadly, for a long time you couldn't use std:: on GPUs.
So, we built a Standard Library for your entire system - CPUs & GPUs
Learn more in my @NVIDIAGTC talk.
Registration is free; watch anytime.
gtc21.event.nvidia.com/media…
1
41
258
If you are:
- A C++, C, or Fortran programmer
- Have Mexican citizenship
- Are interested in participating the C++ committee
Please contact me! Mexico currently doesn't participated in international programming language standardization but I'd like to change that. Please RT.
20
148
245
The C++ Standard Library is not a substitute for a robust package management system.
We cannot standardize ourselves out of this hole.
12
11
237
Three years ago, during my @NVIDIA interview:
"How would you implement `memcpy`"
Now:
"Crap I've got to implement `memcpy`"
9
9
242
I will not be at @CppCon this year because I do not believe it is safe for attendees.
I have great admiration and respect for the volunteers and staff, but leadership has shown that they make bad and secret decisions about attendee safety.
8
19
213
The GPUs are designing themselves now.
Arithmetic circuits were once the craft of human experts, and are now designed by AI in NVIDIA GPUs. H100 chips have nearly 13,000 AI designed circuits! How is this possible? Blog developer.nvidia.com/blog/de… + a thread 🧵👇
9
41
220
`import std` is coming in C++23, and it's FAST - in most cases, faster than including only the headers that you need!
wg21.link/P2412
8
57
222
Over the last few years some of the key figures in C++ twitter have adopted a negative & pessimistic perspective which I find incredibly harmful & hurtful.
I am saddened & disappointed that others have chosen to act this way.
30
19
209
Want to write portable code that can be GPU-accelerated? Come to my #GTC22 talk to find out how you can with Standard C++! I'll cover all the latest developments - senders & receivers, mdspan, and linear algebra.
reg.rainfocus.com/flow/nvidi…
44
213
Attempts to silence me on diversity by "reporting" me to my employer, ISO, or INCITS are pointless.
They all support diversity and inclusiveness in tech.
Likewise, I'm not intimidated by threats against my person.
9
8
190
Stephen Jones, the longtime CUDA architect, is giving a great talk on how to write a CUDA program!
nvidia.com/gtc/session-catal…
2
17
208
44,068
I’ve asked this in interviews at my hospital for 30 years: What medicine do you do at home? Many answer they dont have time. Wrong. Do surgery at home. Every day. If you dont have that passion, being a doctor isn't really for you. Do small surgeries.
I’ve asked this in interviews at my aerospace engineering firm for 30 years: What planes do you build at home? Many answer they dont have time. Wrong. Build missiles at home. Every day. If you dont have that passion aerospace engineering isn't really for you. Launch small rockets
5
36
172
.@nvidia's C++ library team is hiring interns for summer/fall 2022.
You get to work with @ericniebler, @incomputable, @AllisonVacanti, @Guriwesu and many other cool people on the future of Standard C++.
We pay interns generously. Remote is fine.
DM or email me a resume.
7
54
190
We now have open source @nvidia Linux GPU drivers: developer.nvidia.com/blog/nv…
4
25
178
My first month at @NVIDIA:
Me: *something something* UVM.
Them: What's UVM?
Me: Unified Virtual Memory.
Them: Don't call it that, call it Unified Memory, no abbreviation. TLAs are evil.
Me: What's TLA?
Them: Three letter acronym.
5
10
188
In my experience extreme absolute statements like this tend to discredit your point. You can say "I question the future of C++ because the language isn't memory safe" without implying that the work of hundreds of your colleagues is pointless.
14
12
178
The intention here is good, but no. We need to put an end to the rockstar culture in C++. (1/n)
9
22
177
"As one of the architects of CUDA, I hope I can help you write less CUDA code."
I couldn't agree more.
3
9
186
14,669
Can someone recommend wireless headset? Must haves:
Over-ear
Comfortable with glasses
Builtin microphone
Mute button
Price is not an issue.
129
3
166
Meet Grace, NVIDIA's server-class ARM CPU.
This is the processor I wished @NVIDIA when I was still working at @BerkeleyLab.
1
32
165
A few years ago I had to decide my stance on diversity & inclusiveness in C++
I could play it safe, remain silent, & be liked by most of the community
Instead I chose to speak up, knowing it would burn bridges & damage my reputation
I have no regrets about that decision
8
9
152
I'd like to hire you to explain this to the dozens of people replying to me who do not seem to understand this point :)
2
145
NVC++ now supports C++ Standard Parallelism, CUDA C++, OpenACC & OpenMP
They're interoperable - even in the same source file
Now you can combine libraries that use different parallel frameworks
Watch my @NVIDIAGTC talk anytime to learn more. It's free!
gtc21.event.nvidia.com/media…
5
45
161
HPC deserves better parallel compilers. Let me show you one.
Come to my "Inside NVC++ and NVFORTRAN" talk at @NVIDIAGTC tomorrow. We'll be making some exciting announcements.
Registration is free!
gtc21.event.nvidia.com/media…
7
30
158
std::format is one of the best quality of life improvements in C++ in the last decade. Thanks @vzverovich!
2
13
149
My GPA was 1.0/4.0 at one point; I flunked out of college. Today I'm a tech lead at my company and a leader in my tech community.
4 years ago I got a 0 on a quantum physics exam. i met with my professor fearing i needed to change my major & quit physics. today, i’m in a top tier astrophysics Ph.D program & published 2 papers.
STEM is hard for everyone—grades don’t mean you’re not good enough to do it.
3
11
153
.@CppCon leadership and the C++ Foundation Board have violated the community's trust
I stand with @pati_gallardo - on 2021-11-22, she, @gregcons, and I met with the Board, and I told them I must resign @CppCon program chair
Transparency and meaningful change is needed
3
23
138
The logic:
If the President was incapacitated, other states may think our ability to respond to an attack is incapacitated, thus giving them an opportunity.
So, the mobile Strategic Air Command is launched, to remind other states that we are always ready to respond to an attack
11
22
142
Want to learn about CUDA C++ from the people who create it?
Come to our virtual meetup and Q&A on November 22nd 2024! This one will focus on the CUDA Core Compute Libraries, e.g. Thrust, CUB, and libcu++.
github.com/NVIDIA/accelerate…
3
20
155
8,779
Rethink what parts of your code can run on GPUs.
It's a lot more than you may think.
You got parallelism? GPU go brrrrr.
piped.video/playlist?list=PL…
@__simt__ @timothybcosta @code_report
Impressive how many things I'm doing on the GPU instead of the CPU nowadays. This wasn't the case at all just a year ago.
5
34
153
A fair amount of @nvidia GPU hardware docs have recently been open sourced: github.com/NVIDIA/open-gpu-d…
3
50
146
Once is an anomaly.
Twice is a coincidence.
Three times is a pattern.
Four times is enough.
We've had enough of harassment and coverups.
We have failed to make the C++ community safe.
It's time for change. Leadership must prioritize safety.
4
18
135
37,752
If you love in Python, GPUs, and open source, @nvidia is looking for a product manager to drive our Python strategy for HPC and data science! Send me DMs.
2
52
139
Want to learn about CUDA C++? Come to the London C++ meetup tomorrow to hear my talk.
I have some special bonus content as well, only for in person attendees!
meetup.com/cpplondon/events/…
6
10
150
13,002
Adding a borrow checker or similar for C++ or switching to Rust protects code written in the future.
But the issue with C++ isn't future code, it's the billions of lines of existing code.
Solutions like sanitizers that can be applied to existing code are critical.
11
8
142
23,812
You mean after people reacted negatively, he claimed that was what he meant.
His post was toxic and a 30-year veteran should know better, especially in a sector of the industry known for crunch.
2
3
122
Today's my 4 year anniversary at @nvidia. When I started, I was the entirety of the C++ core libraries team; today I'm blessed to have @Guriwesu, @AllisonVacanti, @g_evtushenko, @ericniebler, and many others as coworkers! Thanks to all of you for making this job so rewarding.
6
132