Article "AMD Has Lost"

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

DigDog

Lifer
Jun 3, 2011
13,443
2,084
126
somebody please help me link this because my browser isn't being helpful:


Accoridng to Daniel Newman, principal analyst at Futurum Reseach, AMD "has lost".

I'll directy you straight away to the .. humorous video of Gamers Nexus who are extremely perplexed at the amount of lies and bad facts in said article, to the point where nobody with a bit of knowledge of the current market could make those mistakes. Video starts at 2:40

I really, REALLY hope this isn't a marketing tactic sponsored by Intel because that would .. be disgusting, but also suggest that things are much worse than expected, in Intel.

I hope this is satisfactory.

Iron Woode

Super Moderator
 
Last edited by a moderator:
  • Like
Reactions: spursindonesia

HutchinsonJC

Senior member
Apr 15, 2007
465
201
126
One of the biggest proponents and owners of AMD hardware and a mod on the forums, just a few short years ago had many systems all running Intel, all toward the same goal of curing cancer. We can all see his performance bias for what he uses the systems for. I think you'd be naive to call it out as an AMD bias instead of what it more appropriately is, a performance bias or even a performance per dollar bias.

I don't think AMD has lost. I don't think Intel has lost. AMD finally got a good foot forward and Intel is tripping over a few hurdles. AMD needed that good foot forward as a means to save the company. Intel got complacent in its market leadership position. We ALL can get complacent in doing the same daily work or tasks at times. Sometimes it takes a quick wake up call to bring us back to being and performing our best. AMD got that wake up call, the result of closing in on bankruptcy and Intel is getting that wake up call, the result of competition.
 

SamMaster

Member
Jun 26, 2010
148
75
101

This Toms Hardware article appears to have different numbers from November:

In the conclusion:
1) Intel currently has 242 publicly disclosed vulnerabilities, while AMD has only 16. That’s a 15:1 difference in AMD’s favor. The gap is just too large to ignore.

I'll leave it at that.
 

rbk123

Senior member
Aug 22, 2006
743
345
136
I REALLY hope this isn't a marketing tactic sponsored by Intel

Very very unlikely for a number of reasons. What is likely:
- they are shorting AMD in short term options trading
- they are long Intel in their funds, and probably short term options trading
- probably bought a bunch of Intel positions not too long ago
- he's doubling down on his position that AMD stock will go down, and if he says it long enough, eventually he'll be proven right. No different than those who keep saying the market is overpriced and will go down - which it will, but in the meantime they are getting long-term crushed by index funds
 

yeshua

Member
Aug 7, 2019
166
134
86

Funny how you don't like facts showed in every thread. Maybe because they counter your biased BS.

Want more? How about TomsHardware?

Maybe you should answer that question yourself. Or accuse both Phoronix and TomsHardware of lying.

Please answer these questions just because I need to know if we are on the same grounds here:

1) Do you actually follow Phoronix? (I do, in fact the owner of this website and I correspond quite often)
2) Do you actually run Linux on any of your PCs? (I do, at least four PCs under Linux)
3) Do you know what workflows are real and with are mostly artificial? (I do as I've read all the Phoronix reviews on the topic and I also know how much Phoronix (Michael Larabel) has been misquoted and how his writing style has affected even your own attitude)
4) How many Phoronix reviews of HW vulnerabilities in Intel CPUs have you personally read? (Me? All of them, I've been a Phoronix reader since the inception)
5) Do you understand these vulnerabilities and their applicability? Like what they entail, in what situations they manifest themselves, and in under which circumstances they can be exploited? (I understand)
6) Has any of the vulnerabilities in the Intel CPUs made them "worthless" as indicated in this topic over and over again?
7) Do Spectre-class vulnerabilities affect AMD CPUs?

Here are the latest Phoronix results with somewhat real world tests:


The Intel Core i7 5960X saw 86% the performance out-of-the-box compared to the unmitigated performance, the Core i7 8700K about 84% the unmitigated performance, the Core i5 9400F at 86%, the Core i9 9900K at 88%, and then the new Intel Core i9 10980XE at 97% based upon the tests ran that are affected by Spectre, Meltdown, and other CPU vulnerability mitigations.

I.e. the newest enterprise CPUs are barely affected at all even though they still have the same old "worthless" Sky Lake core.

Also, let's restore some justice here, https://en.wikipedia.org/wiki/Meltdown_(security_vulnerability)#Affected_hardware : the worst vulnerability (by its performance impact) was also found in the CPUs made by ARM, IBM, and probably some older SPARCs. It surely looks to me it wasn't just Intel who sought to make their CPUs faster without thinking about proper security.

Speaking of wildly misinterpreted results in terms of performance loss, here are some perfect examples, i.e. Redis LPUSH, GET and SET benchmarks:


There's almost 38% performance loss. However question number

8) How often do you run Redis LPUSH, GET and SET queries in isolation without having any application logic on top of it? Just to be extra sure, do you know what Redis is? (I do, cause we use it in our company).

Anyways, will please get back to the marketwatch piece and how it's completely irrelevant on these forums? Or people haven't expressed their hatred towards Intel in other topics yet? But are we sure such discussions even pertain to the subforum called "CPUs and OC'ing"? Maybe we should move to politics? ;-)
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,478
14,434
136
One of the biggest proponents and owners of AMD hardware and a mod on the forums, just a few short years ago had many systems all running Intel, all toward the same goal of curing cancer. We can all see his performance bias for what he uses the systems for. I think you'd be naive to call it out as an AMD bias instead of what it more appropriately is, a performance bias or even a performance per dollar bias.

I don't think AMD has lost. I don't think Intel has lost. AMD finally got a good foot forward and Intel is tripping over a few hurdles. AMD needed that good foot forward as a means to save the company. Intel got complacent in its market leadership position. We ALL can get complacent in doing the same daily work or tasks at times. Sometimes it takes a quick wake up call to bring us back to being and performing our best. AMD got that wake up call, the result of closing in on bankruptcy and Intel is getting that wake up call, the result of competition.
Yes, that was me. Until Ryzen came out I would not touch bulldozer., it was 4790k,950 I7, 2600k, 3930k and 6 E5-2683v3 CPU's (and I forget all the rest). First I was only doing Ryzen and Threadripper, now I am stating to build a EPYC farm.

Yes, I am a performance bang/buck advocate, not MFG.
 

VirtualLarry

No Lifer
Aug 25, 2001
56,226
9,990
126
1) Do you actually follow Phoronix? (I do, in fact the owner of this website and I correspond quite often)
Occasionally, although they are a Linux-oriented site, and Linux is not my primary platform
2) Do you actually run Linux on any of your PCs? (I do, at least four PCs under Linux)
I have it installed on a few of my machines, but they are not my daily-drivers.
3) Do you know what workflows are real and with are mostly artificial? (I do as I've read all the Phoronix reviews on the topic and I also know how much Phoronix (Michael Larabel) has been misquoted and how his writing style has affected even your own attitude)
I consider gaming and gaming benchmarks to be about as "Real-world" as you can get, and the article that I quoted, granted, they did a follow-up article indicating that those mitigation patches for the Gen7/Gen7.5 iGPUs were 'preliminary', but the fact that they saw nearly 50% performance loss, belies your comment about "Intel mitigations only causing 10% performance loss in 'real-world' scenarios. Possibly you meant, 'best-case'?
4) How many Phoronix reviews of HW vulnerabilities in Intel CPUs have you personally read? (Me? All of them, I've been a Phoronix reader since the inception)
As many as cross my path
5) Do you understand these vulnerabilities and their applicability? Like what they entail, in what situations they manifest themselves, and in under which circumstances they can be exploited? (I understand)
I like to think that I mostly do. I used to be an x86 ASM programmer back in the 386/486 days.
6) Has any of the vulnerabilities in the Intel CPUs made them "worthless" as indicated in this topic over and over again?
I should have said "For me". As a part-time SOHO system-builder, trying to sell boxes with older Intel CPUs in them, after explaining to them that Intel CPUs have vulnerabilities that can cause someone to steal their SSN or hack their bank account while they're online (in theory), generally makes them not want a system with an Intel CPU in it.
7) Do Spectre-class vulnerabilities affect AMD CPUs?
As far as I am aware, AMD is vulnerable to Spectre V1 exploits, but not Meltdown, or most newer ones, like ZombieLoad.

Regardless, if you were such a "follower of Phoronix", then you should have been explicitly aware of the Gen7/7.5 iGPU Intel vulnerability mitigation taking nearly 50% of performance away (in games), and thus stating that Intel mitigation performance loss was on the order of 10% or less under "real-world" scenarios, was a bit disingenuous.
 

VirtualLarry

No Lifer
Aug 25, 2001
56,226
9,990
126
I.e. the newest enterprise CPUs are barely affected at all even though they still have the same old "worthless" Sky Lake core.
But they're not "the same" (exactly identical), are they? I thought that they had "hardware mitigations" baked-in, eg. slight modifications to the design. Otherwise, if they are just microcode/firmware "Fixes", then Intel is negligent in not back-porting those fixes to prior CPU families.

The flip side of that is, if you are right, and the cores ARE "identical", and the "Fixes" are just microcode, and were back-ported, then the marketing angle that these "new" platforms are "Fixed" from vulnerabilities, is shady as all hell.
 

yeshua

Member
Aug 7, 2019
166
134
86
Regardless, if you were such a "follower of Phoronix", then you should have been explicitly aware of the Gen7/7.5 iGPU Intel vulnerability mitigation taking nearly 50% of performance away (in games), and thus stating that Intel mitigation performance loss was on the order of 10% or less under "real-world" scenarios, was a bit disingenuous.

This result in games is also misquoted.

First of all, this topic is about CPUs, not GPUs (the result you keep referring to).
Second of all, this "fix" has already been reworked and a new proper fix has a much smaller performance impact (less than 10%).

Can you provide some real arguments please? Or we could move to Politics/Economics where we can discuss Intel (mis)management in details. This is a technical subforum where people are expected to have some technical background and first-hand knowledge of the matter but maybe it's just wishful thinking and we are just throwing something at the fan? 😊
 

maddie

Diamond Member
Jul 18, 2010
4,722
4,627
136
Ice Lake has just one more HW vulnerability than AMD: https://en.wikipedia.org/wiki/Transient_execution_CPU_vulnerabilities

They have already fixed eight and I'm quite sure TGL will be comparable to AMD if not better. In no way older Intel CPUs have become worthless and as many tests show in real life scenarios all the mitigations applied mean less than 10% performance loss which can be hardly perceived unless your workflow is running benchmarks all day long. I have two Intel based PCs which work fine and one of them is 13 years old and it's perfectly capable.

I'm going to ask if we have a technical forum here where people are supposed to be knowledgable, unbiased and factual? Then why "Intel == worthless" even though Ice Lake has a better IPC than Zen 2? Why wortless if Zen 2 desktop CPUs have an idle power consumption which is up to ten (!) times higher than of Intel's despite being produced using a much better node? How can a company which invented the x86 instruction set in the first place be called worthless?

Why so many exaggerations and falsehoods?

Again, why are we discussing this useless article from God knows where? Should we create a topic on every inanity spelled on the internet in regard to either Intel or AMD?
Why do I feel I'm in a campaign to reshape my thinking?
 

yeshua

Member
Aug 7, 2019
166
134
86
Why do I feel I'm in a campaign to reshape my thinking?

Whatever suits you sir 😊 I prefer to deal with facts, not something I've read somewhere and keep parroting without even trying to dig deep into the topic. Like poor Phoronix test results which have been misinterpreted/misquoted/overstated by 99% of websites out there because what could be better than proclaiming that, "The Meltdown fix for Intel CPUs destroys their performance" and then read somewhere in the middle of an article that only some of very artificial benchmarks show this level of regression (most tech websites didn't even brother to mention that). The level of tech journalism on most websites is horribly low as yellow headlines are more likely to attract impressions.

Who's this Phoronix guy and how can I join his cult?


Looks like the first hit to me.
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,478
14,434
136
Whatever suits you sir 😊 I prefer to deal with facts, not something I've read somewhere and keep parroting without even trying to dig deep into the topic. Like poor Phoronix test results which have been misinterpreted/misquoted/overstated by 99% of websites out there because what could be better than proclaiming that, "The Meltdown fix for Intel CPUs destroys their performance" and then read somewhere in the middle of an article that only some of very artificial benchmarks show this level of regression (most tech websites didn't even brother to mention that). The level of tech journalism on most websites is horribly low as yellow headlines are more likely to attract impressions.




Looks like the first hit to me.
So the Tomshardware article that states "Phoronix's recent testing of all mitigations in Linux found the fixes reduce Intel's performance by 16% (on average) with Hyper-Threading enabled, while AMD only suffers a 3% average loss. Phoronix derived these percentages from the geometric mean of Linux test results from its test suite." is lying about the results ?

So many members here, and many websites all say Intel gets hit the hardest, and there is a performance slowdown, they are all wrong and you are right ?
 

amrnuke

Golden Member
Apr 24, 2019
1,181
1,772
136
So the Tomshardware article that states "Phoronix's recent testing of all mitigations in Linux found the fixes reduce Intel's performance by 16% (on average) with Hyper-Threading enabled, while AMD only suffers a 3% average loss. Phoronix derived these percentages from the geometric mean of Linux test results from its test suite." is lying about the results ?

So many members here, and many websites all say Intel gets hit the hardest, and there is a performance slowdown, they are all wrong and you are right ?
Mark I think the important piece here is that you have to be specific about which Intel CPU you're testing, and in which timeframe. If you look within the last month at the testing, 10th gen has a 3% performance hit, while 5th gen to 9th gen are seeing a 12-16% hit (biggest hit against the 8th gen on Linux). So I think we have to be very specific. On the server side, Xeon 8280 takes a 5% hit, while 1280v5 takes a 23% hit.

I'm not an Intel apologist, but I think having good information and being granular about what we are talking about is important.

Those headlines can be both right (at the time) and wrong (currently) due to time of publication, data evaluated, etc.
 

RetroZombie

Senior member
Nov 5, 2019
464
386
96
Ice Lake has just one more HW vulnerability than AMD
They have already fixed eight and I'm quite sure TGL will be comparable to AMD if not better.
Then why "Intel == worthless" even though Ice Lake has a better IPC than Zen 2?
The problem is that all intel desktop, HEDT and Server/Datacenter and almost all their mobile cpus are not Ice lake based.

So all your point is completely lost.

Why so many exaggerations and falsehoods?

In no way older Intel CPUs have become worthless and as many tests show in real life scenarios all the mitigations applied mean less than 10% performance loss which can be hardly perceived unless your workflow is running benchmarks all day long.
Are you sure there aren't more incoming?
And I bet many where paid by intel for no disclosure, it all depends on the ethics of the guys that find them...
 

DisEnchantment

Golden Member
Mar 3, 2017
1,590
5,722
136
Regardless, if you were such a "follower of Phoronix", then you should have been explicitly aware of the Gen7/7.5 iGPU Intel vulnerability mitigation taking nearly 50% of performance away (in games), and thus stating that Intel mitigation performance loss was on the order of 10% or less under "real-world" scenarios, was a bit disingenuous.

No need to follow Phoronix, if you are working with Linux on an Intel machine, and running kernel 5.3+ you will notice the performance difference. Compilation for me tanks, VM performance tanks. If you use emulation software like QEMU it tanks hard.
You will have to set up your system to remove all mitigations to regain some performance. Some mitigations can be turned off by kernel boot params. Some mitigations are baked in the kernel so you cant turn off.
If you are a SW dev it sucks.

Who's this Phoronix guy and how can I join his cult?
There is no Phoronix cult, they are called Linux users.
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
Ice Lake has just one more HW vulnerability than AMD:

Um

The problem is that all intel desktop, HEDT and Server/Datacenter and almost all their mobile cpus are not Ice lake based.

Yeah. That.

Very very unlikely for a number of reasons. What is likely:
- they are shorting AMD in short term options trading
- they are long Intel in their funds, and probably short term options trading
- probably bought a bunch of Intel positions not too long ago

If that's true, he may be violating some SEC regulations? Maybe? It would certainly look bad if he came out swinging as a financial analyst on a site like marketwatch trying to manipulate people into defending his positions.

- he's doubling down on his position that AMD stock will go down, and if he says it long enough, eventually he'll be proven right. No different than those who keep saying the market is overpriced and will go down - which it will, but in the meantime they are getting long-term crushed by index funds

Might be the more-likely scenario. For the author's sake, I hope this is the case.
 

yeshua

Member
Aug 7, 2019
166
134
86
The problem is that all intel desktop, HEDT and Server/Datacenter and almost all their mobile cpus are not Ice lake based.

So all your point is completely lost.

Software mitigations have been deployed and work just fine with a negligible performance loss (in most real world scenarios) and then current Cascade Lake/Comet Lake CPUs have most of the reported vulnerabilities fixed in HW, see, https://en.wikipedia.org/wiki/Transient_execution_CPU_vulnerabilities and https://www.intel.com/content/www/u...ngineering-new-protections-into-hardware.html

Are you sure there aren't more incoming?
And I bet many where paid by intel for no disclosure, it all depends on the ethics of the guys that find them...

Are you sure the software you're using is free of vulnerabilities yet to be discovered? Not really, venerabilities are reported and fixed all the time. Does it prevent you from using the said software? No. Has any PC in the world with an Intel CPU become worthless because HW vulnerabilities have been found? No, however the prevalent attitude here on Anandtech is that Intel CPUs are useless even despite the fact that AMD has yet to claim a performance crown under all circumstances. For instance, Intel Core i9 9900K despite being useless/worthless/slowed down to molasses is still faster than all existing AMD CPUs for gaming at 1080p when coupled with high-end GPUs from NVIDIA.

Just to make sure I'm not an Intel shill, my primary rig is based on the Ryzen 3700X CPU.
 

maddie

Diamond Member
Jul 18, 2010
4,722
4,627
136
Software mitigations have been deployed and work just fine with a negligible performance loss (in most real world scenarios) and then current Cascade Lake/Comet Lake CPUs have most of the reported vulnerabilities fixed in HW, see, https://en.wikipedia.org/wiki/Transient_execution_CPU_vulnerabilities and https://www.intel.com/content/www/u...ngineering-new-protections-into-hardware.html



Are you sure the software you're using is free of vulnerabilities yet to be discovered? Not really, venerabilities are reported and fixed all the time. Does it prevent you from using the said software? No. Has any PC in the world with an Intel CPU become worthless because HW vulnerabilities have been found? No, however the prevalent attitude here on Anandtech is that Intel CPUs are useless even despite the fact that AMD has yet to claim a performance crown under all circumstances. For instance, Intel Core i9 9900K despite being useless/worthless/slowed down to molasses is still faster than all existing AMD CPUs for gaming at 1080p when coupled with high-end GPUs from NVIDIA.

Just to make sure I'm not an Intel shill, my primary rig is based on the Ryzen 3700X CPU.
One thing about the net, anyone can pretend to be whomever they want, and it's pretty obvious that you're an Intel fan with an AMD machine. Who knows? In any case, nearly everyone here recommends a top end Intel desktop for high refresh gaming, so I can't see how you can claim that the general recommendation is Intel is worthless, unless thre is something else in play.

What are you really trying to achieve with these posts?
 

yeshua

Member
Aug 7, 2019
166
134
86
One thing about the net, anyone can pretend to be whomever they want, and it's pretty obvious that you're an Intel fan with an AMD machine. Who knows? In any case, nearly everyone here recommends a top end Intel desktop for high refresh gaming, so I can't see how you can claim that the general recommendation is Intel is worthless, unless thre is something else in play.

What are you really trying to achieve with these posts?

I'm sorry, sir. I didn't know that providing strong arguments with a strong knowledge of the matter counts as being an Intel fanboy nowadays. In this case I rest my case. You have won. I'm an Intel fanboy, right. What I tried to achieve could be picked up from my first reply in this "Let's bash Intel one more time" topic. I suggested we stopped discussing some inane news piece on a very dubious website which has little to nothing to do with technology. Again this is a subforum about CPUs and OC'ing, not about companies and their (mis)management.

No need to follow Phoronix, if you are working with Linux on an Intel machine, and running kernel 5.3+ you will notice the performance difference. Compilation for me tanks, VM performance tanks. If you use emulation software like QEMU it tanks hard.
You will have to set up your system to remove all mitigations to regain some performance. Some mitigations can be turned off by kernel boot params. Some mitigations are baked in the kernel so you cant turn off.
If you are a SW dev it sucks.

Let's be honest here, Michael Larabel has shown several times that there's no perceivable performance loss when you're using the mitigations=off parameter and if you only run your own code, or the code you totally trust, then there's no reason not to use it.

I have to agree with you that the companies which provide shared virtual environments (Amazon AWS, Linode, etc.) have been hit hard by these mitigations because they cannot disable them.
 
Last edited:
  • Like
Reactions: CHADBOGA

maddie

Diamond Member
Jul 18, 2010
4,722
4,627
136
I'm sorry, sir. I didn't know that providing strong arguments with a strong knowledge of the matter counts as being an Intel fanboy nowadays. In this case I rest my case. You have won. I'm an Intel fanboy, right. What I tried to achieve could be picked up from my first reply in this "Let's bash Intel one more time" topic. I suggested we stopped discussing some inane news piece on a very dubious website which has little to nothing to do with technology. Again this is a subforum about CPUs and OC'ing, not about companies and their (mis)management.
You've been raging against anyone daring to say Intel is not the best choice at present for most use cases. Trying to shame members for having strong opinions is not a winning strategy to change minds here, assuming that's the intention, as IQs are above average and bullying won't work easily. Even your last line is still trying to suppress criticism. Truly hilarious.
 

nicalandia

Diamond Member
Jan 10, 2019
3,330
5,281
136
This issue is somewhat exacerbated by the fact that some moderators are partial to AMD for some reasons and some people have been banished from these forums because of that which is quite ugly to say the least..
Keep this Up and you will be next so take a chill pill
 
  • Like
Reactions: JPB

RetroZombie

Senior member
Nov 5, 2019
464
386
96
Are you sure the software you're using is free of vulnerabilities yet to be discovered? Not really, venerabilities are reported and fixed all the time. Does it prevent you from using the said software? No. Has any PC in the world with an Intel CPU become worthless because HW vulnerabilities have been found?
The problem is not the worthless, nor i was addressing that with the question, it's the slowdown.
If you prefer:
Are you sure there isn't more intel slowdown incoming?

For example i was really surprised with the major performance hit that the old intel gpu's have with the vulnerabilities, to the point where you go from minimum acceptable into almost worthless?
 

yeshua

Member
Aug 7, 2019
166
134
86
The problem is not the worthless, nor i was addressing that with the question, it's the slowdown.
If you prefer:
Are you sure there isn't more intel slowdown incoming?

For example i was really surprised with the major performance hit that the old intel gpu's have with the vulnerabilities, to the point where you go from minimum acceptable into almost worthless?

https://phoronix.com/scan.php?page=news_item&px=Intel-Recovers-Gen7-Loss
Earlier this month when Intel disclosed CVE-2019-14615 as a security vulnerability affecting their graphics architecture, older Gen7 graphics saw a huge hit to their performance with the initial patches for addressing this vulnerability on Ivy Bridge and Haswell processors. Fortunately, a new mitigation patch series was sent out this week where they believe the performance costs are now avoided.