Jump to content

Talk:Ryzen

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

This is an old revision of this page, as edited by Jeffq (talk | contribs) at 19:50, 31 May 2019 (→What is "Threadripper"?: new section). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

Ryzen is a "brand name" by AMD for microprocessors

Exactly like "Opteron", "Core i3", "Core i5" or "Core i7". I suggest this article should reflect the articles:

and be renamed to List of AMD Ryzen microprocessors

Socket AM4 or Zen (microarchitecture) have their own distinct articles. User:ScotXWt@lk 11:14, 1 March 2017 (UTC)[reply]

Apparently this deserves continued discussion. There have been a number of edits/reverts by KamranMackey titling the Ryzen as a "line" rather than a brand. While line is not defined, AMD has clearly stated that Ryzen is the "brand" for this series of CPUs and APUs.[1] This is supported by reputable secondary sources.[2][3] Further examples are the Athlon and Xeon articles. (An exception, the lead of the AMD Phenom where line is used, is contradicted in the cited source.) Any change in this naming would require a source (and an exceptional one at that to over-rule the manufacturer), otherwise it is original research. Dbsseven (talk) 16:56, 28 July 2017 (UTC)[reply]

References

  1. ^ https://www.amd.com/en-us/press-releases/Pages/amd-takes-computing-2016dec13.aspx
  2. ^ https://venturebeat.com/2016/12/13/amd-introduces-ryzen-brand-for-zen-based-desktop-processors/
  3. ^ http://www.anandtech.com/show/10907/amd-gives-more-zen-details-ryzen-34-ghz-nvme-neural-net-prediction-25-mhz-boost-steps

http://ark.intel.com/ equivalent?

Is there such a database by AMD? User:ScotXWt@lk 11:52, 1 March 2017 (UTC)[reply]

There doesn't seem to be. 119.92.137.221 (talk) 08:53, 12 January 2018 (UTC)[reply]
https://wikichip.com/ and https://specdb.info/ -- are the closest things to Ark neither site run by amd themselves. Specdb is the closest to Ark, ran by a reddit user. Wikichip is significantly more technical but IIRC amd employees do add to them. --81.97.194.212 (talk) 02:12, 14 November 2018 (UTC)[reply]

Impact

I think this article should be a List and nothing else, but until it is, how about documenting the impact of Ryzen-branded microprocessors:

Intel's x86-64 competition:

I would compare this article to Kaby Lake. Do we even have confirmation that Zen is still the official architecture name, or is the architecture now Ryzen?Carewolf (talk) 15:12, 1 March 2017 (UTC)[reply]
I guess at some point this article could be made an overview Ryzen branded CPUs, and the current content moved to a Summit Ridge article as it is specific to that implementation of Zen. I am just not certain AMD is certain how they are going to brand this in the long run.Carewolf (talk) 15:22, 1 March 2017 (UTC)[reply]
The impact is rather simple. AMD offers more value for the same money than what Intel has done for many years, consequently Intel gradually matches this to some extend and also offers more cores for the same price. Intel had a monopoly for many years and consequently Intel offered little value for money (contrasted to the dieshrinks and lower costs for the same number of cores), AMD brought the competition back and consequently in the next years you will see much more progress. I agree that there should be a section dedicated to this aspect of Ryzen, though I would put it in the wiki on Zen because Ryzen is just the first iteration of Zen and Zen is the general architecture (in the X86-family). One final note, because of the big impact Ryzen should definitely get a higher importance rating on the hardware scale. It is much more than just a new architecture, it is a small revolution for CPUs in this decade.
145.132.75.218 (talk) 18:26, 6 March 2018 (UTC)[reply]

pronounce

Any hint how to pronounce? Thanks — Preceding unsigned comment added by 86.59.222.27 (talk) 00:36, 3 March 2017 (UTC)[reply]

The podcasts PC Perspective and This week in computer hardware say Ry-Zen. Dbsseven (talk) 18:53, 17 March 2017 (UTC)[reply]
I don't think the current IPA pronunciation hits the nail on the head. "Ryzen" as in "(Ho)rizon" would be spelt /'raɪzɘn/, which is how AMD pronounces this. /'raɪzɛn:/ would make it two individually pronounced syllables as in Rye-Zen, but again, AMD says it differently.--Aera (talk) 23:10, 24 July 2017 (UTC)[reply]
A year on and nobody is pronouncing it "Rye-zen". There are numerous video clips in which AMD staff say the word as "Rye-zn" with the e being hardly vocalised at all. As it stands the article is incorrect. 83.104.249.240 (talk) 02:44, 27 April 2018 (UTC)[reply]

Product lineup table

Simple enough edit to make, but the "threadripper" row in the table doesn't have enough columns, as such the "memory support" column reads "Summer 2017", and the Release price has no entry at all.


As a wikipedia newbie, I thought this would be a simple enough edit to make, to get me started on my way Find a load of "unknown"s and copy and paste one more, plus its delimiter. Simple! Instead all I see it say is: AMD Ryzen inside a pair of double braces (I'm not going to type that literally just in case it has some special meaning in the talk page that I don't know about. Now I thought that meant hyperlink, so when I went to that page. It just redirects me back to this article.


It's a minor low-priority cosmetic edit at this point, I'm sure it will be fixed when all the details are known, but until then, it'd make the page look a bit better if it was fixed. While I get to feel a little silly for not knowing how to edit something that seems like it should be really simple. 88.98.86.251 (talk) 20:46, 19 May 2017 (UTC)[reply]

I already fixed it. You can find the table by going to the Template:AMD Ryzen page. It's a template, so the table can be re-used on multiple pages. Thanks for contributing! Dbsseven (talk) 20:58, 19 May 2017 (UTC)[reply]
Ah! "Template", that's the bit I didn't understand, I thought it was just a link to a standard page, and why I got confused when it just redirected right back where I came from. Like I said, it seemed like it should've been something really simple, and I bet now that I know I'm looking for the word template, I'll find it straight away in the help page about tables... 88.98.86.251 (talk) 16:51, 20 May 2017 (UTC)[reply]

add phonetic spelling for Ryzen

many non amarican people are pronouncing it wrong. Please at ITA for it — Preceding unsigned comment added by 91.17.202.236 (talk)

Wikipedia is a wiki, you can edit it yourself. But anyway I've added a pronounciation. I don't think many people understand ITA, I think "rye-zen" is clearer. -- intgr [talk] 21:54, 30 May 2017 (UTC)[reply]
Except that isn't how it's pronounced. It's more like "rye-zn". 83.104.249.240 (talk) 02:46, 27 April 2018 (UTC)[reply]

Intel ADX

http://cdn.wccftech.com/wp-content/uploads/2016/08/AMD-Zen_Instruction-Set-ISA-740x416.png

We should add this to the list of extensions that Ryzen supports. --135.0.52.247 (talk) 01:31, 6 June 2017 (UTC)[reply]

product Line up targeted Market change suggestions

I am pretty new to wiki. When i first looked at this article a few months ago, Ryzen 7 was an "High-End" Processor, but now it is classified as an "Performance" Processor. I find this kinda confusing and wrong, because Ryzen 7 is still an High-End CPU, while Threadripper is an Enthusiast-class CPU on an Enthusiast-Plattform. This should be changed on the graph in this article. It only adds confusion in my opinion. I would like to edit the table, but i dont know how to edit it. Additionally i want to hear the opinion of u guys, before i make any changes.

Here are my suggestions:

  • AMD Ryzen Threadripper: Enthusiast
  • AMD Ryzen 7  : High-End
  • AMD Ryzen 5  : Performance (Ryzen 5 1600X, Ryzen 5 1600), Mainstream (Ryzen 5 1500X, Ryzen 5 1400)
  • AMD Ryzen 3  : Entry Level

— Preceding unsigned comment added by AMD-User (talk • contribs)

Interesting ideas. However, the categories listed are based on current citations per Wikipedia policy. If you have a citation for your proposed edits, feel free to post them here to try and find a new consensus. Dbsseven (talk) 17:19, 22 June 2017 (UTC)[reply]
I strongly disagree to add more crap from "AMD marketing". These people should be fired anyway... We should have articles on concrete die/chip packages or a family of thereof, e.g. Clarkdale (microprocessor), on microarchitectures, e.g. AMD K8 (though the article itself is severely lacking), and yes, the Wikipedia could continue to bother with lists of …-articles, e.g. List of Nvidia graphics processing units or List of AMD FX microprocessors. User:ScotXWt@lk 07:40, 27 March 2019 (UTC)[reply]

Switch to binary units?

Should we maybe switch to binary units where applicable? After all, "AMD defines 1 kilobyte (KB) as 1024 bytes, and 1 megabyte (MB) as 1024 kilobytes", while by definition 1KB is 1000B, but 1KiB do match AMD's 1024B. Ceremony64 (talk) 08:12, 3 August 2017 (UTC)[reply]

@Ceremony64: I think Wikipedia manual of style indicates "MB" and "KB" here. See WP:COMPUNITS "The IEC prefixes kibi- (symbol Ki), mebi- (Mi), gibi- (Gi), etc., are generally not to be used except: [...]" -- intgr [talk] 08:23, 3 August 2017 (UTC)[reply]

Binary units would be better. After all, all units in computer science are a product of 2 because a switch (transistor, in this case) can be on or off. 145.132.75.218 (talk) 06:59, 7 March 2018 (UTC)[reply]

Product lineup table outdated

Someone needs to update the info in Product lineup table, which seems to contain outdated info. For example Ryzen 1700X, clock rate and memory speed listed doesn't match the specs in official page https://www.amd.com/en/products/cpu/amd-ryzen-7-pro-1700x Hei Liebrecht 19:05, 2 November 2017 (UTC)[reply]

Good catch. Updated. Checking on the pro products, it looks like only the Pro 1700X speeds were wrong. Dbsseven (talk) 19:36, 2 November 2017 (UTC)[reply]

Regarding XFR it seems officially AMD just says whether it's available, but not how much speed it gives. Since Pro seems like a re-label of the orignal Ryzen CPUs, would it be safe to assume that the XFR speed of Pro and original models match?--Pizzahut2 (talk) 20:27, 2 November 2017 (UTC)[reply]

I'd rather not assume, personally. I presume there is/will be some source to publish this. Dbsseven (talk) 00:23, 8 November 2017 (UTC)[reply]

infobox core name

I realize this is a bit detailed but the "Core names" of the infobox is not correct. However, the listed Summit Ridge, Whitehaven, and Raven Ridge are product code names not core names. The core name is "Zen". The infobox template also includes "Product code" and "Model", might not one of these be more precise? Dbsseven (talk) 00:15, 8 November 2017 (UTC)[reply]

Basically just look at pretty much all the articles about Intel products, and do it exactly the same! Most articles related to AMD products are "cr**", ehm, severely lacking. AFAIK a "code name" is not the same as a "brand", but in AMD articles these are used inconsistently… Why???? User:ScotXWt@lk 07:43, 27 March 2019 (UTC)[reply]

Ryzen brand logo vs actual picture of an AM4 Ryzen 1800X for article's primary image?

Pretty much what it says in the subject line; I just saw that someone recently swapped out the articles primary image from being a rendering of the AMD Ryzen logo to an actual picture of a Ryzen chip; specifically the 1800X. Now I'm of the belief that the former picture suited the article much better than the new/current one for a number of reasons. First off, not every Ryzen chip uses the AM4 socket so showing a picture of an AM4 chip as the primary could be very misleading. Secondly this article is about the entire Ryzen brand of chips, not a single product or set of chips whithin that brand, thus I feel like returning to the brand logo would be the best header picture possible. Just wondering people's thoughts on the matter before I just go and do it though. — Preceding unsigned comment added by 2605:A601:30A6:9D00:98E1:1D3A:F3CE:3EF6 (talk) 07:24, 11 December 2017 (UTC)[reply]

I suspect the reasoning is that the photo is freely licensed, while use of the logo relies on fair use or whatever terms AMD licenses it under. Since the image isn't showing the underside of the chip, it isn't really sufficient to determine the socket type alone, so I don't think that is likely to confuse people. And the image is captioned with the exact model it portrays, which should remove any confusion. --James (talk) 08:47, 8 January 2018 (UTC)[reply]
I personally think that the Ryzen logo is preferable, licensing permitting. 119.92.137.221 (talk) 08:52, 12 January 2018 (UTC)[reply]

FPUs

Is it correct that 6 cores do have integrated FPUs each, not sharing them? 80.109.210.78 (talk) 13:21, 3 January 2018 (UTC)[reply]

Yes. This is discussed on the Zen (microarchitecture) article and links within that article. Dbsseven (talk) 14:49, 3 January 2018 (UTC)[reply]

Desktop APUs

Should the recently announced and soon to be released 2200G and 2400G be added to the main table of products (Ryzen 3, 5, 7 and Threadripper) or should they have a new table of their own? If the former then the table will need modifications. Clearly they are not Mobile Ryzen products but a new table with similar features could be created. 119.92.137.221 (talk) 09:02, 12 January 2018 (UTC)[reply]

I began a discussion on the Template:AMD Ryzen talk page about this very topic, in an effort to avoid the template becoming unwieldily. But yes, this content should certainly be added. Dbsseven (talk) 15:37, 12 January 2018 (UTC)[reply]
I'll take a look there. 119.92.137.221 (talk) 02:10, 13 January 2018 (UTC)[reply]

Whitehaven?

In the Summit Ridge/Whitehaven subsection the term Summit Ridge is defined but Whitehaven is not. I assume Whitehaven refers to first generation Threadripper but I'm reluctant to add it without a citation. 119.92.137.221 (talk) 02:08, 13 January 2018 (UTC)[reply]

Duplication

The product tables are being duplicated between this page and the Zen microarchitecture page. Where does it belong best? It looks like AMD will be sticking with the Ryzen brand for a good long time like they did with the Atlhon brand. Should we move products to their architecture page and have a simple listing of iterations here (list features, not specific products)? Real tlhingan (talk) 21:03, 30 January 2018 (UTC)[reply]

The product tables are defined as templates so they only exist in one place but can be included in as many articles as need them. 83.104.249.240 (talk) 06:31, 12 February 2018 (UTC)[reply]
I think Real tlhingan's point is that the content is redundant, and it may not be necessary to replicate it. Based on previous pages (Intel and AMD) it is consensus to replicate the tables over both the microarchitecture and product pages. (ie. Core i7 and Kaby Lake) I personally find this very helpful, and with the templates it is easy to keep the content consistent. Dbsseven (talk) 14:12, 12 February 2018 (UTC)[reply]
Zen+ Ryzen went retail. So, we add Ryzen 2000 here as a new table and copy to Zen+? So Ryzen would consist of all Ryzen generation and each architecture only related generations? Elk Salmon (talk) 02:28, 8 April 2018 (UTC)[reply]
Zen+ is a refresh of Zen and should not have an article all to itself. Curly "JFC" Turkey 🍁 ¡gobble! 06:02, 8 April 2018 (UTC)[reply]

Too technical?

I severely disagree with the statement that this article would be too technical and I suggest that we don't make it less technical. If it gets difficult for laymen to understand parts of it, then explain those parts in more detail. Technology is difficult, get over it! We don't dumb down wikis on mathematics, let's not do that for technical subjects either. On top of that is Ryzen a new kind of architecture within the X86-family and consequently it gets a bit more complicated. I have read the entire article, it really isn't that difficult to understand most of it, nothing in this article is in-depth. Having said that, I do think that this article could be improved, I notice that it could use a lot more structure. — Preceding unsigned comment added by 145.132.75.218 (talk) 18:19, 6 March 2018 (UTC)[reply]

Agreed, especially on the comparison with the often incomprehensibility of mathematical articles. The article does need more content and explanation. 83.104.249.240 (talk) 16:52, 24 April 2018 (UTC)[reply]

Coverage of CTS Labs vulnerabilities

@Cautilus, Denniss, and Jasonanaggie: I'm reverting the addition again for now. I agree that these are real vulnerabilities and worthy of coverage on Wikipedia. But they need to be covered neutrally: independent security researchers agree that there is nothing "critical" about these vulnerabilities, there is no "backdoor", etc. and AMD is already working with hardware vendors to distribute fixes.

PS: In the future when there's a disagreement, please create a talk page section to discuss it; edit summaries are not the right place to have a discussion. See WP:BRD. -- intgr [talk] 15:24, 23 March 2018 (UTC)[reply]

softlock bug

@Denniss: and IP editor(s): I am beginning this discussion to see if consensus can be found about the purported GNU/linux soft lock bug of the Ryzen processors. Right now the citations provided appear (to me) not to meet the standards for reliable sources (based on WP:SPS) and and for content (based on WP:NOTCHANGELOG). Unless and until there is a reliable third party publication discussing these, I don't think we have evidence they are encyclopedic and therefore warrant inclusion. Thoughts? Dbsseven (talk) 23:19, 17 May 2018 (UTC)[reply]

Everything was fine until some contributors started removing one command/variable claiming "it's only used for testing". So the problem appeared with certain Ryzen system. Users adding it back have no problem so it's clearly a software problem as something in the AMD ecosystem under Linux relied on it. If I read the bugreport correct in kernel 4.12 all was fine, only with 4.13+ the problem appeared due to removal of this command/variable. Reminds me of "never change a running system" as there was no real need to remove it.--Denniss (talk) 06:22, 18 May 2018 (UTC)[reply]

Ryzenfall Tag discussion - possible questions and help?

I just tagged the Ryzenfall section for bias, orig research, and sources. I came across it doing research on Intel vs AMD vulnerabilities and noticed the tone of the section seemed off immediately and so checked the source links. I don't have the time to really source this correctly myself right now, nor am I an expert on the subject.

I'd like to discuss the sources I have problems with: One of the sources (currently listed as cite 134: Lawyer Blog) is a personal blog on a crowd sourced, non-reviewed, article aggregator, which is clearly out. Another is less clear cut, (cite 133, Gamer Nexus) because it's a medium sized youtube channel and gamer site. The article is written by one of it's main Editors and an assistant, there are no other real published accomplishments for either in their field - so their expertise is debatable. The article itself is not sourced except for inline mentions, no links or references, and highly slanted with original research and fringe theories - the former is poor material and the latter is, of course, allowable, but amusing. The most borderline reference in the problem area is what is currently cite 132 (Linus Torvalds) which is a post on the Linus Torvalds forums with quotes and references from various, likely legitimate, 2nd and 3rd party sources, often with Linus Torvalds direct quotes. Torvalds is directly invested in software that had to make patches and updates for some of the vulnerabilities found between Intel and AMD and I'm unsure if he would count as a "third party" reference who is too close to the issue. That is the issue that may prevent simple switching of the link to one of the other potentially more reliable sources listed in the forum post?

Any feedback would be appreciated - or if you're a more informed expert on this subject and can de-clutter, show some more relevant data, or neutralize the tone - even better. I don't really even know where to start looking for good security sources for up-to-date news. — Preceding unsigned comment added by Thumbcat (talk • contribs) 23:44, 15 August 2018 (UTC)[reply]

To be honest, I'm not sure how to keep neutrality in this topic. The topic very quickly came and gone, and not much was revealed throughout the entire thing- CTS acted shrewdly, AMD was quite quiet through the entire controversy, with the first message saying that they received the reports but have never heard of CTS, and the second message saying that they are working on a fix. RAM (talk) 03:51, 20 August 2018 (UTC)[reply]
The whole thing was mainly a scam. The section does list both sides, but it could probably be considered undue weight. Though the story had coverage enough to warrant mention somewhere on wp.Carewolf (talk) 19:00, 20 August 2018 (UTC)[reply]
IMHO its not really even an errata, it's more so some very lenient security checks as the assumption is attacker won't have admin and they won't have physical access and you won't download random firmware, not from AMD.
Things like the Phenom TLB bug, Spectre, Meltdown, L1TF, Pentium Fdiv. are all key examples of major errata that 1) has a major impact to computing 2) affects the product significantly where it could be no longer fit for purpose 3) had to be worked around.
The "errata" demoed by CTS had requirements like Ring 0 execution or custom firmware/bios to be installed. I guess it could be more suited to Trivia if we don't want it removed? with just what CTS claimed, Media coverage and AMDs response.
And if we were to consider this an errata where do we draw the line as there's another 46 known errata(s) for Ryzen according to the AMD 17h Rev Guide and around 175 of known errata to Skylake according to the 6th Gen Spec Update. 81.97.194.212 (talk) 02:43, 14 November 2018 (UTC)[reply]

ECC Support?

Is it true that the Series 2 Ryzens have dropped support for ECC memory?

I noticed a marked increase in crashes of my current (INTEL) PC when I utilised non ECC memory and my PC does need to be highly available. My new one (hopefully Ryzen) as it will work both as freeview recorder and Network video recorder.

If so this is a significant change for some types of system. Processing ECC 'off CPU' will create a one cycle performance hit. --Leopardtail (talk) 08:41, 4 September 2018 (UTC)[reply]

No. It isn't true. But show me an AM4 motherboard that has both hardware support and BIOS support for ECC RAM. Note that a number of AM4 motherboards claim that ECC RAM can be used but that is very different from saying that the ECC function can be used. 83.104.249.240 (talk) 19:21, 6 November 2018 (UTC)[reply]

Raven Ridge number of PCIe lanes

The article claims "16 external PCIe 3.0 lanes (4 to chipset). 8 internal PCIe 3.0 lanes for iGPU." I added a Dubious-Discuss tag because Raven Ridge has 16 internal PCIe 3.0 lanes for the iGPU plus other on-board I/O, such as USB and SATA. Since Zen supports a total of 32 PCIe lanes (with a maximum of 24 available externally) that leaves 16 available externally for Raven Ridge Ryzens, as follows: 4 to the Promontory chipset, 4 to the M.2 socket, 8 to the dGPU x16 slot. This can be shown using tools such as HWiNFO64 but doesn't seem to be documented anywhere. The Raven Ridge Athlon 200GE has fewer PCIe lanes available - only 4 to the dGPU slot and (possibly) no support for MVMe M.2 devices. 83.104.249.240 (talk) 19:43, 6 November 2018 (UTC)[reply]

The Zepplin dies have 32 PCIe lanes. AM4 only allows for a total of 24 of them to be used. TR4/SPr3 will use the full 32 lanes. Raven Ridge is a different chip to Zepplin, it has a total of 16 PCIe lanes only has 8 PCIe lanes after the 8 have been used by the iGPU 81.97.194.212 (talk) 17:32, 15 November 2018 (UTC)[reply]
Something is wrong with your maths there. Raven Ridge still has 4 lanes to the Promontory chip and 4 lanes to the M.2 socket, so (with the 8 lanes to the dGPU slot) that's 16 lanes total at the AM4 socket. That doesn't account for the internal connections. 83.104.249.240 (talk) 03:49, 17 February 2019 (UTC)[reply]

FMA3 Errata

Why on earth someone wants to just call this errata, give it a specific name. Errata is a general term. Spectre is a form of errata, Meltdown is a form of errata. --81.97.194.212 (talk) 17:11, 15 November 2018 (UTC)[reply]

These are vulnerabilities, not errata. An erratum is a correction for a published text.--84.130.207.138 (talk) 12:38, 12 January 2019 (UTC)[reply]

Proposed merge

Someone proposed that the article List of AMD Ryzen microprocessors be merged into this article but neglected to create a section to discuss it here. So I'll make a start.

I don't see any information in that article that isn't already present in this article, except for the list of Epyc processors. Since Epyc is not Ryzen it should not be there anyway. So my vote is not to merge but to delete the other article. 83.104.249.240 (talk) 04:00, 17 February 2019 (UTC)[reply]

Probably better to just split off the processor information into the "list of" article, like what happened with Qualcomm Snapdragon.198.52.130.167 (talk) 23:07, 31 March 2019 (UTC)[reply]

Compatibility/Issues - are these sections really needed and neutral?

I think these sections are a bit odd. Long since 1985 (386) AMD archs have been consistently running x86 code perfectly fine, yet the myth of "AMD compatibility" still persists. Intel had many serious bugs: https://en.wikipedia.org/wiki/Pentium_F00F_bug https://en.wikipedia.org/wiki/Pentium_FDIV_bug https://groups.google.com/forum/#!topic/nairobi-gnu/EQcYcRZa9Os https://www.eetimes.com/document.asp?doc_id=1125791 https://en.wikipedia.org/wiki/Sandy_Bridge#Cougar_Point_chipset_flaw https://en.wikipedia.org/wiki/Intel_Atom#Issues https://www.extremetech.com/computing/244074-intel-atom-c2000-bug-killing-products-multiple-manufacturers https://en.wikipedia.org/wiki/Intel_Management_Engine#Security_vulnerabilities https://en.wikipedia.org/wiki/Spectre_(security_vulnerability)#Impact https://en.wikipedia.org/wiki/Meltdown_(security_vulnerability)#Affected_hardware https://arxiv.org/pdf/1903.00446.pdf

Meanwhile AMD just couldn't run code compiled with strict optimizations for proprietary post-586 Intel archs: a notable example were past versions of OpenFOAM, an opensource fluid dynamics package. It managed to have the TLB bug https://en.wikipedia.org/wiki/AMD_Phenom#Background which wasn't observable in real life, initial Ryzen bugs have been fixed by firmware updates and has been involved in Spectre/Meltdown much less than Intel.

point 1: Ryzen(and more generally, "Zen") arch is x86 compatible, or is there some evidence of the contrary to justify the voice Compatibility? point 2: All chips have bugs, yet Intel Core arch wikipage https://en.wikipedia.org/wiki/Intel_Core doesn't have a "Bugs" section, but Ryzen does sport an "Issues" section. This looks like a subtle way of sowing distrust or fud. I think every company should have the same key sections. Cheavacca (talk) 14:27, 27 February 2019 (UTC) edited 10/03/2019 to add Spoiler exploit[reply]

I'm not sure what your complaint is regarding the Compatibility section. There's nothing debatable in there at all, it's completely factual and well sourced.
Regarding point 2, whether Intel artices have an Issues section is completely irrelevant. Rather than a conspiracy, the more likely explanation is that no one has had the time add such sections to the articles. I'm sure you're more than welcome to do so yourself.
As for your claim that initial Ryzen bugs have been fixed by firmware updates, this is false. Bugs that could be patched in firmware have indeed been patched, but the Linux segfault bug is a hardware issue and requires the processor to be replaced. --Veikk0.ma 04:15, 12 March 2019 (UTC)[reply]
@Cheavacca: I haven't heard any "myths of AMD compatibility" myself. Also the section seems fairly netural, it states that Microsoft doesn't officially support Intel CPUs with older Windows versions either.
I don't see any reason to remove the section but you have any thoughts about how to make the section more neutral then that would be welcome.
We could also rename "Compatibility" to "Operating system support". -- intgr [talk] 14:26, 12 March 2019 (UTC)[reply]
@intgr: you must be very young then… maybe stop talking about "myths" and ask people about their experience with concrete products from the 386/486-era. Also please distinguish between compiler and operating system. I don't know how many people even use Intel's compiler, but they do bother to offer one. Though, do they still? User:ScotXWt@lk 07:49, 27 March 2019 (UTC)[reply]
Yeah I'm too young to know about 386/486 issues, but that's irrelevant. We're talking about Ryzen in particular here, Cheavacca suggested that the myths still persist; I haven't heard of compatibility myths or issues about Ryzen or any modern x86 processor -- have you?
As for compilers, "Compatibility" section in the acticle currently only discusses operating system support (e.g. ), not compilers. I don't think Intel's compiler was ever truly incompatible with AMD processors, it just disabled many optimizations that would be applicable. -- intgr [talk] 09:53, 27 March 2019 (UTC)[reply]

Proposed merge.

I think the articles need their own spaces. These graphs on the list are well-done, but would clog the main Ryzen page. 70.18.10.58 (talk) 09:36, 8 April 2019 (UTC)[reply]

Ryzen Embedded R1000 series released!

Hello guys

A month ago AMD released a new series in its Embedded section. The Ryzen R1000 series, which consists of 2 models, R1505G and R1606G, were released during a keynote in Taipei last month. I tried to edit the page to include that but the site returned me with HTTP 404. Can someone else help edit that so this newest information can also be added? Check all the specs out here: https://www.amd.com/en/products/specifications/embedded/11411

Thank you so much.

What is "Threadripper"?

Allthogh Threadripper reirects here, that term is not defined in any way by this current article. Onr might be able to infer that "Threadripper" is AMD's term for a high-end processor, it's neither stated clearly nor described meaningfully. I ask that regular editors of this article add a brief statement in the lede to clarify this ambiguity. Thank you. ~ Jeff Q (talk) 19:50, 31 May 2019 (UTC)[reply]

Retrieved from "https://en.wikipedia.org/w/index.php?title=Talk:Ryzen&oldid=899705916"