- 103
- 79 824
CSC-ECE 506: Architecture of Parallel Computers
เข้าร่วมเมื่อ 11 ส.ค. 2020
Lecture 11d. Replacement policies and mechanisms
Lecture 11d. Replacement policies and mechanisms
มุมมอง: 462
วีดีโอ
Lecture 11c. Inclusive, exclusive, and NINE caches
มุมมอง 1.3K3 ปีที่แล้ว
Lecture 11c. Inclusive, exclusive, and NINE caches
Lecture 11b. Virtually vs. physically indexed caches
มุมมอง 1.1K3 ปีที่แล้ว
Lecture 11b. Virtually vs. physically indexed caches
Lecture 11a. Translation lookaside buffers
มุมมอง 5673 ปีที่แล้ว
Lecture 11a. Translation lookaside buffers
Lecture 1a. Course objectives & policies (summer)
มุมมอง 1K3 ปีที่แล้ว
Lecture 1a. Course objectives & policies (summer)
Lecture 1a. Course objectives and policies
มุมมอง 4943 ปีที่แล้ว
Lecture 1a. Course objectives and policies
Lecture 16d. Load linked/Store conditional (LL/SC)
มุมมอง 2.4K3 ปีที่แล้ว
Lecture 16d. Load linked/Store conditional (LL/SC)
Lecture 19a. Relaxed memory-consistency models
มุมมอง 1.4K3 ปีที่แล้ว
Lecture 19a. Relaxed memory-consistency models
Lecture 19b. Sequential and causal consistency
มุมมอง 13K3 ปีที่แล้ว
Lecture 19b. Sequential and causal consistency
Lecture 20e. Alternatives for organizing directories
มุมมอง 1793 ปีที่แล้ว
Lecture 20e. Alternatives for organizing directories
Lecture 16c. Test and test and set lock TTSL
มุมมอง 4613 ปีที่แล้ว
Lecture 16c. Test and test and set lock TTSL
Lecture 11d. Replacement policies and mechanisms
มุมมอง 373 ปีที่แล้ว
Lecture 11d. Replacement policies and mechanisms
Lecture 20a. How to scale a multiprocessor
มุมมอง 3773 ปีที่แล้ว
Lecture 20a. How to scale a multiprocessor
Lecture 20b. Bus-based vs. directory-based coherence
มุมมอง 2233 ปีที่แล้ว
Lecture 20b. Bus-based vs. directory-based coherence
Lecture 17a. Centralized barrier implementations
มุมมอง 3903 ปีที่แล้ว
Lecture 17a. Centralized barrier implementations
Lecture 17b. Distributed barrier implementations
มุมมอง 1233 ปีที่แล้ว
Lecture 17b. Distributed barrier implementations
Lecture 18a. The two hypotheses of memory consistency
มุมมอง 5263 ปีที่แล้ว
Lecture 18a. The two hypotheses of memory consistency
Lecture 18b. Sequentially consistent outcomes
มุมมอง 3793 ปีที่แล้ว
Lecture 18b. Sequentially consistent outcomes
a video on MOESI please
Only (b) is deadlock-free.
Gracias thanks
Really appreciate the clear & concise way this was presented.
Hi, there is typo on 5:43, when a P1 is in state Sm, and it snooped a BusRd, it doesn't update the main memory, only P3. But top-left of the screen, it says P1 flushes, sending update to P3 and main memory.
There is a typo/mistake at 2:00, it says that "if there is a cpu write miss and the shared line is asserted, the changes is written through to main memory, and the block is cached in state D." According to the diagram, in this case, the block will be cached in S, not D.
great explanation
This is fantastic. Thanks for posting.
In 6:40, why W(x)1 is not causally related to R(x)1, as you said before? does not make sense.
Thanks for the video! It helped tremendously in understanding multilevel cache policies!
From E to S and I, when BusRd or BusRdx occurs, why Flush is needed. Can't requesting processor directly read it from main memory?
the best video explaining MSI, thank you
You are welcome. Thanks also to Yan Solihin, who made the original slides, and Mohit Gambhir, who helped put everything together.
The part about writes not reaching all processes instantly was what I was missing. Thanks
pRDON DA BEN MATLAB GÖRMÜYORUM GERİZEKALI DEVREDE GÖSTERSENE
Perfect Presentation
So clear Explanation.Thanks a lot !!!
Nice, finally understood causal consistency!
Great explanation! Thank you :)
great explanation
Your voice sounded so familiar, and today I figured it out, it's the AI in 2001 space odyssey.
u sure??
Thanks
Please upload more lectures
Awesome lecture professor. Thanks from india
what does "BusRdX" mean?
BusRd exclusive. You are reading a block into the cache, and you need to assure that yours is the only processor that has it cached, so that you can write it without causing data to become incoherent.
Thank you for the explanation
Hi, just wanted to say that this was really usefull and well explained. Thanks
Thank you so much. This is the only good explanation of release consistency that I could find on TH-cam.
Good video, that helped a lot!