Real World Tech
@rwt
Followers
1K
Following
12
Media
4
Statuses
119
The official twitter feed for https://t.co/mK6sJo4O1d.
San Francisco, CA
Joined February 2011
Ashlee is a friend and an amazing writer! You won’t be disappointed with his book :)
1
1
5
Someone loves our forums!!!
Linus on AVX512: https://t.co/POhaucigvH "[…] yes, compilers have gotten a lot better, but not only did vectorization turn out to be much harder in practice than the trivial examples were […] I personally cannot find a single benchmark that does anything I would ever do […]"
0
0
6
Intel 4 achieves 20% better performance, scales logic density 2X, while reducing costs through design co-optimization, new materials, and judicious use of EUV lithography. It will ramp with the Meteor Lake compute tile to high volume manufacturing in 2023
realworldtech.com
The Intel 4 process achieves 20% better performance and scales logic density by 2X while reducing costs through extensive design co-optimization, adoption of new materials, and judicious use of EUV...
4
23
97
Intel just released a fantastic history of the transistor, tracing the evolution of this marvelous technology. The dimensions are accurate too :)
0
5
15
If there are any RWT readers attending #hotchips2021 we have a slack channel for real time chat - drop in to h-rwt and chat with .@TheKanter and others!
0
1
1
Happy 25th birthday to me! It’s been a wonderful quarter century of articles and forum discussions…here’s to 25 more!!!
0
2
20
The Hotchips 2021 program is live! Check it out at https://t.co/aQ840fG91f Congrats to all the amazing presenters! We can't wait to hear the great talks. See you Aug 22-24, don't forget to register at
hotchips.org
A Symposium on High Performance Chips
0
22
51
We have upgraded our VM, so everything should be faster now!!! Enjoy :)
1
1
12
Happy anniversary to the whole RWT community, which is now 24 years old! We are thrilled to still have lively discussions and occasional articles and hope to continue for many years to come!
0
3
14
Vendors brag about the transistor count of their design. However, transistor count, density varies based on the design. The result is that transistor count, density are approximate metrics and focusing on those numbers loses sight of the bigger picture
realworldtech.com
Transistor count and transistor density are often portrayed as technical achievements and milestones. Many vendors brag about the complexity of their design, as measured by transistor count. In...
1
6
18
Power delivery is one of the most significant challenges in modern processors. In this article we discuss system power delivery goals and techniques, the impact on frequency, #Intel FIVR, decoupling capacitors and other topics
realworldtech.com
Power delivery is one of the most significant challenges in modern processors. The power delivery network (PDN) must meet the demanding requirements of modern CMOS technology, supply power with...
1
7
18
Power delivery is one of the most significant challenges in modern processors. In this article we discuss system power delivery goals and techniques, the impact on frequency, #Intel FIVR, decoupling capacitors and other topics
realworldtech.com
Power delivery is one of the most significant challenges in modern processors. The power delivery network (PDN) must meet the demanding requirements of modern CMOS technology, supply power with...
1
7
18
The theme of #SC19 was combining machine learning with HPC. We heard about new technologies from Nvidia, Graphcore, and Cerebras Systems and the challenges of benchmarking HPC systems for machine learning.
realworldtech.com
For me, SC19 was about the fusion of machine learning and scientific computing. I learned about new technologies from Nvidia, Graphcore, and Cerebras Systems and spoke on a panel about the role of...
0
0
2
Speaking with David Kanter about how RDNA is actually different from GCN with the AMD Navi GPUs, including an explanation of waves, warps, and cache:
8
22
135
We are upgrading our VM today, so expect a little downtime while we reboot!
0
0
0
Congratulations to the #MLPerf community (including @TheKanter) who helped create the v0.5 machine learning training benchmarks. Results are public, with great work by Google, Intel, and Nvidia.
1
1
8