Redfire75369’s Twitter Archive

Popular Tweets

A list of popular tweets by retweets and favorites.

1. Semiconductor Capital Expenditure 2022: Pureplay Foundries: TSMC: 40-44B UMC: 3.0B GFS: 4.5B SMIC: 5.0B Memory Foundries: Micron: 11-12B SK Hynix: >11.12B Integrated Device Manufacturers: Intel: 25-28B Texas Instruments: 3.5B STMicro: 3.4-3.6B Infineon: 2.4B
   Permalink On twitter.com ♻️ 41 Retweets ❤️ 157 Favorites 2022 Feb 11 Mood 0
2. High resolution Alder Lake-S Die Shot: I believe this was sent to reviewers and it was on embargo till today. I'm not 100% sure. Resolution: 6950x3514 Link: 1drv.ms/u/s!ApJwjdCrzFEchWhNOJjFVlO3bab3

   /img/RfYWt31_rP.jpeg
   Permalink On twitter.com ♻️ 21 Retweets ❤️ 75 Favorites 2021 Nov 4 Mood +1 🙂
3. There's a bunch of interesting pictures here of future products. Sapphire Rapids/HBM 1. Package 2. Substrate 3. Tiles cnet.com/pictures/a-look-inside-intels-mammoth-arizona-chipmaking-fab/

   /img/Yzz_K5fNCj.jpeg/img/g8duOpgD7V.jpeg/img/ao-NaefsM3.jpeg
   Permalink On twitter.com ♻️ 18 Retweets ❤️ 70 Favorites 2021 Nov 18 Mood +2 🙂
4. Process Node Table/Graph Update: The major differences are regarding Intel 4, TSMC N6/N5, Samsung 4LPE and derivatives therein. There's a lot of other changes, but you'll have to compare them manually for that.

   
   Permalink On twitter.com ♻️ 17 Retweets ❤️ 90 Favorites 2022 Jul 4 Mood 0
5. We now have pictures of the Ponte Vecchio package without the tiles on it, similar to Sapphire Rapids-HBM. 1. Ponte Vecchio Package 2. Sapphire Rapids-HBM Package nytimes.com/2022/04/08/technology/intel-chip-shortage.html

   
   Permalink On twitter.com ♻️ 16 Retweets ❤️ 79 Favorites 2022 Apr 10 Mood 0
6. Data Center GPU Comparison H100 vs MI250X vs PVC

   
   Permalink On twitter.com ♻️ 14 Retweets ❤️ 62 Favorites 2022 Mar 22 Mood 0
7. So... TSMC has settled on a 3-year cadence between N3 and N2. TSMC is probably the most complacent leading-edge logic foundry at this point I scoffed when Samsung said 2GAE was in 2025 and I'll do it again with TSMC as they've said N2 is in H2-2025.
   Permalink On twitter.com ♻️ 13 Retweets ❤️ 105 Favorites 2022 Apr 15 Mood -2 🙁
8. Frontier's MI250X is higher clocked as compared to those for the "regular market". They're clocked at 1.88 GHz for a total of 53.0 TFLOPs FP64 per OAM. For comparison, the regular MI250X is clocked at 1.7 GHz for a total of 47.9 TFLOPs FP64 per OAM.
   Permalink On twitter.com ♻️ 10 Retweets ❤️ 40 Favorites 2022 May 5 Mood 0
9. Process Node Table/Graph Update: TSMC N3 density from 1.7× to 1.6× relative to N5. (293 to 274 MTr/mm²). TSMC N7 density from 96.5 to 91.3 MTr/mm². Samsung 2GAE in 2025. Updated colours on graph to reflect company's brand colours.

   
   Permalink On twitter.com ♻️ 10 Retweets ❤️ 66 Favorites 2021 Nov 9 Mood 0
10. My summary of Intel Architecture Day 2021 is finally complete! redfire75369.github.io/summaries/events/intel/architecture-day-2021
    Permalink On twitter.com ♻️ 9 Retweets ❤️ 24 Favorites 2021 Aug 26 Mood 0
11. Nvidia Data Center GPU MCM Approach: Composable On-Package Architecture (COPA) Divergence of HPC and DL GPUs by modifying I/O and cache, depending on the workload, using chiplets. (1/x)

    
    Permalink On twitter.com ♻️ 9 Retweets ❤️ 48 Favorites 2021 Dec 14 Mood 0
12. …in reply to @Redfire75369
    Meteor Lake 1. Wafer 2. Test Chip 3. Test Chips 4. Packaging

    
    Permalink On twitter.com ♻️ 9 Retweets ❤️ 51 Favorites 2021 Nov 18 Mood 0
13. Sapphire Rapids MCC Mesh Diagram: This is my interpretation of the die layout based on the die shot we have gotten. If you find anything wrong, bring it up. I've probably stared at this for too long and gotten tunnel vision.

    
    Permalink On twitter.com ♻️ 8 Retweets ❤️ 46 Favorites 2022 Sep 28 Mood -7 🙁
14. …in reply to @david_schor
    @david_schor For some context: equipmenthandbooks.com/2012/06/liebherr-lr-13000-on-crane-days/

    
    Permalink On twitter.com ♻️ 8 Retweets ❤️ 13 Favorites 2022 Jul 9 Mood 0
15. WSJ: TSMC Looks to Build Multibillion-Dollar Chip Plant in Singapore I highly, HIGHLY doubt this... wsj.com/articles/tsmc-looks-to-build-multibillion-dollar-chip-plant-in-singapore-11652958840
    Permalink On twitter.com ♻️ 8 Retweets ❤️ 35 Favorites 2022 May 19 Mood -1 🙁
16. Of course, noone is going to call out AMD for this, so I guess I will. If it's +33% to +40%, why is your bar 3x taller? This is quite literally Intel-level marketing.

    
    Permalink On twitter.com ♻️ 6 Retweets ❤️ 64 Favorites 2021 Nov 8 Mood 0
17. With the Biren and new Intel disclosures on their HPC/AI Data Center GPUs, the table of peak compute throughputs has been updated. I also added Xeon Phi, PEZY, NEC and Preferred Networks for no reason.

    
    Permalink On twitter.com ♻️ 6 Retweets ❤️ 50 Favorites 2022 Aug 22 Mood -1 🙁
18. Today is my birthday! 🎂🎉⚡🎉🎂 My only wishes are for competition to keep raging and for WATTS to keep rising in the year ahead!

    
    Permalink On twitter.com ♻️ 4 Retweets ❤️ 80 Favorites 2022 Mar 22 Mood +1 🙂
19. With 10000 nodes, Aurora has 2.80 EFLOPs of peak FP64 performance. That's 43MW purely on the CPU and GPU.
    Permalink On twitter.com ♻️ 4 Retweets ❤️ 55 Favorites 2022 May 10 Mood 0
20. AMD Navi III (RDNA 3) Navi 31: 12288 SPs, 48 WGPs, 12 SAs, 6 SEs Navi 32: 8192 SPs, 32 WGPs, 8 SAs, 4 SEs Navi 33: 4096 SPs, 16 WGPs, 4 SAs, 2 SEs t.co/h46ytpdifo
    Permalink On twitter.com ♻️ 4 Retweets ❤️ 54 Favorites 2022 May 2 Mood 0
21. …in reply to @nixcraft
    @nixcraft Holy C, of course, how could it be anything else?
    Permalink On twitter.com ❤️ 53 Favorites 2021 Nov 7 Mood 0