BRN 0.00% 24.5¢ brainchip holdings ltd

Ann: BrainChip to present at TechKnow Invest Roadshow, page-15

ANNOUNCEMENT SPONSORED BY PLUS500
ANNOUNCEMENT SPONSORED BY PLUS500
CFD TRADING PLATFORM
CFD Service. Your Capital is at risk
CFD TRADING PLATFORM CFD Service. Your Capital is at risk
ANNOUNCEMENT SPONSORED BY PLUS500
CFD TRADING PLATFORM CFD Service. Your Capital is at risk
  1. 1,554 Posts.
    lightbulb Created with Sketch. 403
    Activity

    Hiren Majmudar Linkedin

    2nd degree connection2nd
    Vice President & Managing Director, Intel Capital

    Austin, Texas

    ConnectConnect with Hiren MajmudarMessageSend a message to Hiren MajmudarMore actionsMore…

    947 followers

    • View Hiren’s profile
      Compute in memory with both DRAM and SRAM approaches are feasible and being researched by promising folks. Obviously one is difficult to manufacture/scale w.r.t. computing vs. other given current manufacturing process of choice.
      Hiren replied to a comment
    • View Hiren’s profile
      With barrage of announcements coming out of AI HW Summit last week, here are my thoughts on where the compute innovations are progressing. Silicon innovation, valuations, VC$ for Silicon are all at an all time high. #CPU #GPU #FPGA #machinelearning #chiparchitecture 1. CPUs will support more inference because they still have headroom of free cycles; GPUs, with added power, will also support more inference. CPUs and GPUs are here to stay (ease of programming and general purpose)! 2. FPGAs are flexible but industry has to solve the cost & ease of programming issues. Innovation towards ease of programming for FPGA will likely have outsized returns. 3. ASICs offer an order of magnitude $$ & Energy efficiency but are less flexible than FPGA. ASICs allow more compute parallelism, ability to optimize memory bandwidth and hence better for emerging “purpose driven architectures”. There will be few winners in this race. 4. Compute in memory is about to become real, will improve latency significantly; Photonic computing may be the next frontier of innovation. 5. Too many open source FWs (for NN models) are painful for HW providers. Standardization is needed. Starting to see early signs of this. 6. Too many benchmarks for showing the value of silicon for AI.
      Hiren shared this
      11 Likes 2 Comments

    • Congratulations DavidD & Avigdor
      Hiren commented

    • Inference on Edge innovation continues. Spiking Neural Network SNNs with a type of neuromorphic architecture from Brainchip.
      Hiren shared this
      5 Likes
 
watchlist Created with Sketch. Add BRN (ASX) to my watchlist
(20min delay)
Last
24.5¢
Change
0.000(0.00%)
Mkt cap ! $483.2M
Open High Low Value Volume
24.5¢ 25.5¢ 23.0¢ $2.156M 8.797M

Buyers (Bids)

No. Vol. Price($)
3 63333 24.0¢
 

Sellers (Offers)

Price($) Vol. No.
24.5¢ 99875 4
View Market Depth
Last trade - 16.10pm 04/11/2024 (20 minute delay) ?
BRN (ASX) Chart
arrow-down-2 Created with Sketch. arrow-down-2 Created with Sketch.