Seems AMD is launching around low end processor line: https://www.guru3d.com/articles-pages/amd-ryzen-3-3100-and-3300x-review,1.ht... Its 120 dollars US for 4 cores/8 threads. According to my research seems to be equal to a i7700K and those go used for around 300 US. May be a interesting option for a budget system if your workloads aren't thread heavy. Seems Intel was really just sitting around after Sandy Bridge for the most part, Nick -- Fundamentally an organism has conscious mental states if and only if there is something that it is like to be that organism--something it is like for the organism. - Thomas Nagel
AMD did a major re-architecture, with the server "chips" having four 120-thread actual chips communicating through a system controller, just like an old parallel mainframe. This all fit in one chip carrier: I'd have loved to see how the interconnects worked, but they were pretty closed-mouth about that They did admit the first generation gave up about 15% when you made two of these coordinate with one another via a cache-consistency chip. I'm waiting to hear about the second generation I suspect the new low-cost chip is a simplified version of one of the four processor cores, and might cost about 1/5 of what the server chips cost to make. --dave On 2020-05-08 10:55 p.m., Nicholas Krause via talk wrote: Seems AMD is launching around low end processor line: https://www.guru3d.com/articles-pages/amd-ryzen-3-3100-and-3300x-review,1.ht... Its 120 dollars US for 4 cores/8 threads. According to my research seems to be equal to a i7700K and those go used for around 300 US. May be a interesting option for a budget system if your workloads aren't thread heavy. Seems Intel was really just sitting around after Sandy Bridge for the most part, Nick -- David Collier-Brown, | Always do right. This will gratify System Programmer and Author | some people and astonish the rest dave.collier-brown@indexexchange.com<mailto:dave.collier-brown@indexexchange.com> | -- Mark Twain CONFIDENTIALITY NOTICE AND DISCLAIMER : This telecommunication, including any and all attachments, contains confidential information intended only for the person(s) to whom it is addressed. Any dissemination, distribution, copying or disclosure is strictly prohibited and is not a waiver of confidentiality. If you have received this telecommunication in error, please notify the sender immediately by return electronic mail and delete the message from your inbox and deleted items folders. This telecommunication does not constitute an express or implied agreement to conduct transactions by electronic means, nor does it constitute a contract offer, a contract amendment or an acceptance of a contract offer. Contract terms contained in this telecommunication are subject to legal review and the completion of formal documentation and are not binding until same is confirmed in writing and has been signed by an authorized signatory.
On Sat, May 09, 2020 at 08:00:33AM -0400, Dave Collier-Brown via talk wrote:
I suspect the new low-cost chip is a simplified version of one of the four processor cores, and might cost about 1/5 of what the server chips cost to make.
Or rather, ones that failed QA. :-) -- William Park <opengeometry@yahoo.ca>
| From: William Park via talk <talk@gtalug.org> | | On Sat, May 09, 2020 at 08:00:33AM -0400, Dave Collier-Brown via talk wrote: | > I suspect the new low-cost chip is a simplified version of one of the | > four processor cores, and might cost about 1/5 of what the server | > chips cost to make. | | Or rather, ones that failed QA. :-) Why simplified? Why Failed? Each Zen 2 package is made from a group of "chiplets". Each processor chiplet (CCX) has 4 cores, with each providing 2 SMT units: 4c 8t in current notation. AMD uses a 7nm TSMC process to make them. Then they have a chiplet or several to do I/O. Density doesn't matter as much here so they are made using something like a 14nm process. So: a low-end Zen 2 package has fewer chiplets. It probably has a lower clock speed. AMD might "bin" them for speed. That means: manufacture them, test how fast they can run, and sell the faster ones as separate SKUs.
On 2020-05-09 3:17 p.m., D. Hugh Redelmeier via talk wrote:
| From: William Park via talk <talk@gtalug.org> | | On Sat, May 09, 2020 at 08:00:33AM -0400, Dave Collier-Brown via talk wrote: | > I suspect the new low-cost chip is a simplified version of one of the | > four processor cores, and might cost about 1/5 of what the server | > chips cost to make. | | Or rather, ones that failed QA. :-)
Why simplified? Why Failed?
Each Zen 2 package is made from a group of "chiplets". Each processor chiplet (CCX) has 4 cores, with each providing 2 SMT units: 4c 8t in current notation. AMD uses a 7nm TSMC process to make them.
Then they have a chiplet or several to do I/O. Density doesn't matter as much here so they are made using something like a 14nm process.
So: a low-end Zen 2 package has fewer chiplets. It probably has a lower clock speed.
That's what I was thinking of when I said "simplified". --dave -- David Collier-Brown, | Always do right. This will gratify System Programmer and Author | some people and astonish the rest dave.collier-brown@indexexchange.com | -- Mark Twain CONFIDENTIALITY NOTICE AND DISCLAIMER : This telecommunication, including any and all attachments, contains confidential information intended only for the person(s) to whom it is addressed. Any dissemination, distribution, copying or disclosure is strictly prohibited and is not a waiver of confidentiality. If you have received this telecommunication in error, please notify the sender immediately by return electronic mail and delete the message from your inbox and deleted items folders. This telecommunication does not constitute an express or implied agreement to conduct transactions by electronic means, nor does it constitute a contract offer, a contract amendment or an acceptance of a contract offer. Contract terms contained in this telecommunication are subject to legal review and the completion of formal documentation and are not binding until same is confirmed in writing and has been signed by an authorized signatory.
On Sat, May 09, 2020 at 08:00:33AM -0400, Dave Collier-Brown via talk wrote:
I suspect the new low-cost chip is a simplified version of one of the four processor cores, and might cost about 1/5 of what the server chips cost to make. Or rather, ones that failed QA. :-) Its nice to see 4 cores become mainstream. A lot of programs don't take advantage of the extra threads if you look into it. Actually most desktop programs don't scale pass 2 fast cores including video games. With 4 cores being mainstream and at a low
On 5/9/20 1:16 PM, William Park via talk wrote: price it will be nice to see software especially for the desktop start to be written more with 4 and more cores in mind. Hopefully this does happen but I'm assuming it will take a few years or so. Through I would mention the extra 20 dollars on the 3100x is worth it due to it being one module if you know about AMD'S CCX interconnects. Its similar to NUMA but on one chip for those who don't know. Cheers, Nick -- Fundamentally an organism has conscious mental states if and only if there is something that it is like to be that organism--something it is like for the organism. - Thomas Nagel
On Sat, 9 May 2020 at 17:42, Nicholas Krause via talk <talk@gtalug.org> wrote:
be nice to see software especially for the desktop start to be written more with 4 and more cores in mind. Hopefully this does happen but I'm assuming it will take a few years or so.
Four cores and seven years ago... :-) -- Scott
On 2020-05-09 5:41 p.m., Nicholas Krause via talk wrote: Its nice to see 4 cores become mainstream. A lot of programs don't take advantage of the extra threads if you look into it. Actually most desktop programs don't scale pass 2 fast cores including video games. With 4 cores being mainstream and at a low price it will be nice to see software especially for the desktop start to be written more with 4 and more cores in mind. Hopefully this does happen but I'm assuming it will take a few years or so. In an earlier life, we noticed a huge increase in usability going from a single core to a dual-core processor in desktop SPARCs. Having two processors available to respond to a request means way more that you'd expect in responsiveness. Somewhat underpowered workstations felt as if they were big multiprocessors. Joyce noticed, and bought me a "Tadpole" UltraSPARC-1c laptop for a birthday present. --dave -- David Collier-Brown, | Always do right. This will gratify System Programmer and Author | some people and astonish the rest dave.collier-brown@indexexchange.com<mailto:dave.collier-brown@indexexchange.com> | -- Mark Twain CONFIDENTIALITY NOTICE AND DISCLAIMER : This telecommunication, including any and all attachments, contains confidential information intended only for the person(s) to whom it is addressed. Any dissemination, distribution, copying or disclosure is strictly prohibited and is not a waiver of confidentiality. If you have received this telecommunication in error, please notify the sender immediately by return electronic mail and delete the message from your inbox and deleted items folders. This telecommunication does not constitute an express or implied agreement to conduct transactions by electronic means, nor does it constitute a contract offer, a contract amendment or an acceptance of a contract offer. Contract terms contained in this telecommunication are subject to legal review and the completion of formal documentation and are not binding until same is confirmed in writing and has been signed by an authorized signatory.
participants (5)
-
D. Hugh Redelmeier -
Dave Collier-Brown -
Nicholas Krause -
Scott Allen -
William Park