"Intel in jeopardy" from insiders. Much much worse than what they tell.

To eleborate a little, lets take a little example, lets say we build a Rack POD for VMware/HyperV/ESXi ect.

Lets say it looks like this:

1 x 1U server
16 x 2 U servers
2 x FC SW
2 x QFX SW

When we use the SAME servers (Mobo,CPU, RAM etc.) we can move the entire farm to another POD with no issues.
But if we start mixing hardware....oh boy here comes the nightmare, man-hours and pain.
When you runs 100's of such POD's...you do not go with the warm feelings of forum-posters, that only know their own little PC...or a single generation of benches.
For us to go AMD...Intel would have to make EPIC mistakes...because of AMD's yoyo history in enterprise...this is smart thinking...we usually try and plan minimum 3 years ahead...in order to save us from pains like the one I just described.

Again, this is not liltte Jimmy's gaming PC we are talking about.
Intel has a proven track-record...AMD does not.
It is not rocket science...
 
delita said:
You and everyone else. I've been trying to push my customers to AMD due to the shortage for months now and they'd rather wait on increasingly longer wait times than switch to something they aren't comfortable with. And this isn't even just on the server side, a lot of time it scales right into the client side which, lets be honest, the average user wouldn't know what kind of processor they had in their work issued notebook. But people have this stigma about AMD for some reason, despite you showing them graphs, talking about the shortage that likely isn't going anywhere any time soon. It's definitely interesting.
Click to expand...

Read my latest reply...it might give you some insight ;)
 
ND40oz said:
If you're on VMware, EVC modes are either Intel or AMD, so vMotion won't work at all between platforms. You'll need completely new clusters of all AMD hosts to switch over and you'll need shut down the VMs to do so. It's not exactly just a trivial change of just hardware.
Click to expand...

Thanks for the informative answer....glad I am not the only one spotting the issues ;)
 
Factum said:
To eleborate a little, lets take a little example, lets say we build a Rack POD for VMware/HyperV/ESXi ect.

Lets say it looks like this:

1 x 1U server
16 x 2 U servers
2 x FC SW
2 x QFX SW

When we use the SAME servers (Mobo,CPU, RAM etc.) we can move the entire farm to another POD with no issues.
But if we start mixing hardware....oh boy here comes the nightmare, man-hours and pain.
When you runs 100's of such POD's...you do not go with the warm feelings of forum-posters, that only know their own little PC...or a single generation of benches.
For us to go AMD...Intel would have to make EPIC mistakes...because of AMD's yoyo history in enterprise...this is smart thinking...we usually try and plan minimum 3 years ahead...in order to save us from pains like the one I just described.

Again, this is not liltte Jimmy's gaming PC we are talking about.
Intel has a proven track-record...AMD does not.
It is not rocket science...
Click to expand...

I definitely appreciate this information, being on an enterprise sales team with a very large national MSP. I certainly don't always have the answers, but what you're saying here makes a lot of sense.
 
delita said:
I definitely appreciate this information, being on an enterprise sales team with a very large national MSP. I certainly don't always have the answers, but what you're saying here makes a lot of sense.
Click to expand...

You are welcome! :)
You have to consider what replacing the underlying architechture does for normal operations....operations people are paranoid by nature (and for good reasons).
And then you have to find a way to mitigate the history of AMD in enterprise...I know people in the industry that took the plunge to AMD with "Operton"...only to get burned a few years later...and have to go back to Intel.
I would wager you would never get them to try again...unless they had NO other option....simply to much hassle over the longer time scale.
 
Factum said:
To eleborate a little, lets take a little example, lets say we build a Rack POD for VMware/HyperV/ESXi ect.

Lets say it looks like this:

1 x 1U server
16 x 2 U servers
2 x FC SW
2 x QFX SW

When we use the SAME servers (Mobo,CPU, RAM etc.) we can move the entire farm to another POD with no issues.
But if we start mixing hardware....oh boy here comes the nightmare, man-hours and pain.
When you runs 100's of such POD's...you do not go with the warm feelings of forum-posters, that only know their own little PC...or a single generation of benches.
For us to go AMD...Intel would have to make EPIC mistakes...because of AMD's yoyo history in enterprise...this is smart thinking...we usually try and plan minimum 3 years ahead...in order to save us from pains like the one I just described.

Again, this is not liltte Jimmy's gaming PC we are talking about.
Intel has a proven track-record...AMD does not.
It is not rocket science...
Click to expand...

I'll just respond with a quote from the conclusion to ServeTheHome's EPYC 7002 series overview (link):
Customers need to adopt AMD EPYC. To our readers, it is important when you get a quote to at minimum quote an AMD EPYC alternative on every order. More important, follow through and buy ones where Intel is not competitive. If AMD EPYC 7002, with a massive core count, memory bandwidth, PCIe generation and lane count, power consumption, and pricing advantage cannot take significant share, we are basically done. If AMD does not gain enormous share with this much of a lead, and easy compatibility, Intel officially has a monopoly on the market and companies like Ampere and Marvell should shut down their Arm projects. If AMD does not gain significant share, there is no merit to having a wholistically better product than Intel.
Click to expand...

Choosing familiarity and comfort over progress is likely to have a lasting negative impact.
 
BrotherMichigan said:
I'll just respond with a quote from the conclusion to ServeTheHome's EPYC 7002 series overview (link):


Choosing familiarity and comfort over progress is likely to have a lasting negative impact.
Click to expand...

That is not AMD's major problem.
AMD's history of being a yoyo is their problem in enterprise.
They have NEVER proved staying power...so would would I create more work for myself down the line?
It takes a sustained effort over YEARS for them to even become an option I might consider.
AMD's problem in enterprise is AMD's history in enterprise...and that is not the fault of Intel, but AMD's own fault.
 
"If AMD does not gain significant share, there is no merit to having a wholistically better product than Intel."

That AMD has a wholistically better product is an opinion.

As much as I'd agree that they do, my perspective ranges only from single-user to very small enterprise deployments that do not yet leverage virtual environments.

In contrast, at the scale that Factum is discussing, well, it's not hard to imagine that AMD still has quite some distance to go. Enterprise sales aren't just spec-sheet fencing.


Now, as a counter-perspective, I see AMD as having two advantages here that they didn't have with Opterons:
  • Intel is having supply issues -- meaning that time-sensitive purchases may go to AMD by default
  • The growth of cloud infrastructure that is purposefully as hardware agnostic as possible provides demand for whatever hardware to fit the bill, and AMD fits that bill very well
It should just be noted that while AMD is winning at racing benchmarks as well as price vs. performance, uptake of their products in commercial settings will likely not track as nearly as close to update by consumers as outsiders would probably expect.
 
I'm not commenting as to the motives of different purchasers, rather just the impact that will have on future growth and development of the market. A widespread hitching of the pants and putting in a little more time and effort to support a better product (incentivizing competition, as it were) seems like it can only be a good thing in the long run.
 
Well I will say its not the responsibility of the consumer to buy product X to save it. Its the responsibility of the manufacturer to engage me enough so that I want to spend my money. Intel is just failing lately. Its thier fault if they are hurting. I will never feel guilt for the decisions others make.
 
tangoseal said:
Well I will say its not the responsibility of the consumer to buy product X to save it. Its the responsibility of the manufacturer to engage me enough so that I want to spend my money. Intel is just failing lately. Its thier fault if they are hurting. I will never feel guilt for the decisions others make.
Click to expand...

From a enterprise perspective Intel is not failing...they have a HUGE deman...bigger than they can fill currently...they sell EVERYTHING they make...and could sell more....that is some "crisis"...
 
It's kind of a Catch-22. The industry needs people to support AMD when they're competitive so they can keep supporting the market and spurring innovation, but people won't buy them because they didn't have the resources or the product to maintain a competitive position in the past.
 
BrotherMichigan said:
I'm not commenting as to the motives of different purchasers, rather just the impact that will have on future growth and development of the market. A widespread hitching of the pants and putting in a little more time and effort to support a better product (incentivizing competition, as it were) seems like it can only be a good thing in the long run.
Click to expand...

If that vendor is not going to be relevant in 3-5 years...that would be enough of a deterrent to stay the course.
I know my sentiment is shared by many in the industry...we care about the looooong run...and Intel would have to make an EPIC fail and AMD have to be present for YEARS before that changes.
 
BrotherMichigan said:
It's kind of a Catch-22. The industry needs people to support AMD when they're competitive so they can keep supporting the market and spurring innovation, but people won't buy them because they didn't have the resources or the product to maintain a competitive position in the past.
Click to expand...

They have a long and hard road ahead of them for sure...becuase they have popped up before...and then dropped away again.
That doesn't entice enterprise one tiny bit.
So we will watch...for some years...and see if anything is diffrent this time...or the yoyo history repeats itself all over again...as before.
 
Not sure if someone already posted on this.

Intel has a new undisclosed security problem on their CPU until Coffee Lake and not sure how Ice Lake wouldn't be affected either.
It's called CACHEOUT.
Intel is already working on this since it has been undisclosed and even DARPA already into it for some time.
It's an extension of Zombieload at the core of the problem.
This is huge, of the kind of Meltdown and Spectre, maybe bigger.
You need to deactivate Hyperthreading and TSX and also make some change in the BIOS (Intel not ready for microcode distribution to motherboard manufacturers).
Intel CPUs predating to Skylake are unaffected and are affected up until Comet Lake (yet to see). In fact all Intel CPU conceived before Q4 2019.
Windows and Linux kernel needs to be patched.
ARM and IBM Power may or may not have a similar problem because they use a function with the same behavior. Not yet tested.

AMD CPUs are unaffected !


https://www.techspot.com/news/83754...cacheout-attack-targets-intel-processors.html

https://www.tomshardware.com/news/intel-responds-to-zombieload-and-cacheout-attacks
 
Funny how some enterprise users say AMD migration is a little tricky but they are still doing it or at least giving it a crack, whilst others don't seem to want to do any of that work and act like ostritches.
AMD of 2020 isn't AMD of 2004 anymore and software support is improving across the board. But it still seems to be an emotion-based decision making process for some.
 
N4CR said:
Funny how some enterprise users say AMD migration is a little tricky but they are still doing it or at least giving it a crack, whilst others don't seem to want to do any of that work and act like ostritches.
AMD of 2020 isn't AMD of 2004 anymore and software support is improving across the board. But it still seems to be an emotion-based decision making process for some.
Click to expand...

You don't seem to understand - Intel has been forever and will be forever. they can deliver flawed trash but because they are intel - it's ok. Now AMD needs to be in enterprise space for only a couple more decades and execute flawlessly with a better product. Then maybe the old timer IT guys will try a couple AMD servers. Or - maybe the payments from Intel are flowing again.
 
Jandor said:
Not sure if someone already posted on this.

Intel has a new undisclosed security problem on their CPU until Coffee Lake and not sure how Ice Lake wouldn't be affected either.
It's called CACHEOUT.
Intel is already working on this since it has been undisclosed and even DARPA already into it for some time.
It's an extension of Zombieload at the core of the problem.
This is huge, of the kind of Meltdown and Spectre, maybe bigger.
You need to deactivate Hyperthreading and TSX and also make some change in the BIOS (Intel not ready for microcode distribution to motherboard manufacturers).
Intel CPUs predating to Skylake are unaffected and are affected up until Comet Lake (yet to see). In fact all Intel CPU conceived before Q4 2019.
Windows and Linux kernel needs to be patched.
ARM and IBM Power may or may not have a similar problem because they use a function with the same behavior. Not yet tested.

AMD CPUs are unaffected !


https://www.techspot.com/news/83754...cacheout-attack-targets-intel-processors.html

https://www.tomshardware.com/news/intel-responds-to-zombieload-and-cacheout-attacks
Click to expand...

None off those things has ever affected our setup, nor will it ever...in fact I have never heard of a successful attack in "the wild"?

N4CR said:
Funny how some enterprise users say AMD migration is a little tricky but they are still doing it or at least giving it a crack, whilst others don't seem to want to do any of that work and act like ostritches.
AMD of 2020 isn't AMD of 2004 anymore and software support is improving across the board. But it still seems to be an emotion-based decision making process for some.
Click to expand...

It is based on ~30 years of experience...going to take a LONG to to rectify that...again, this is not little Jimmy's gaming PC.

But funny that I can tell who works in enterprise and who are a "little Jimmy gamer" based solely on their postings here...
 
30 years of the Intel salesman taking the executive IT staff out to the strip clubs, Yeah, the Oracle guy had the same Corporate credit card to pay the tab. Little jimmy stayed at work babysitting and patching Intel boxes.
 
Last edited:
Factum said:
None off those things has ever affected our setup, nor will it ever...in fact I have never heard of a successful attack in "the wild"?



It is based on ~30 years of experience...going to take a LONG to to rectify that...again, this is not little Jimmy's gaming PC.

But funny that I can tell who works in enterprise and who are a "little Jimmy gamer" based solely on their postings here...
Click to expand...

Ok boomer.
 
N4CR said:
Funny how some enterprise users say AMD migration is a little tricky but they are still doing it or at least giving it a crack, whilst others don't seem to want to do any of that work and act like ostritches.
AMD of 2020 isn't AMD of 2004 anymore and software support is improving across the board. But it still seems to be an emotion-based decision making process for some.
Click to expand...

It's not the initial amount of work to migrate an instance, it's more work but it's also introducing systems that are completely incompatible with the rest of the environment. I'm just migrating test and dev work loads off of HP DL380 G6s, systems that have been in use for almost 10 years. Everything else in the VMware environment moves down a rung and the new boxes take over the highest tier production workloads. So if I switch over, nothing in my environment can migrate in and out of the highest tier cluster without shut downs. Sorry not happening, I'd need to replace all of my hardware at once, all of my environments to switch over to AMD. It doesn't make any sense to do that, fiscally or level of effort.
 
vick1000 said:
Dollar for dollar, AMD wins, pound for pound Intel wins.
Click to expand...


{sarcasm}The Exchange rate after Brexit isn't going to be that good!!!! If it means intel CPU's are going to be cheaper here in the UK then I am all for it LOLOLOLOLOL {/sarcasm}
 
As someone who really only cares about gaming performance, I'm waiting for AMD to release a CPU that can actually beat Intel in single threaded gaming applications. Sure, Intel's a disaster right now, but AMD's been such a disaster they still haven't caught up with Intel's single thread performance, which is pretty pathetic, honestly.

Also, I'm sick of worthless benchmarks. I don't care about average framerates. I want to see minimum framerates.

If the averages between two CPUs are 105fps and 110fps, but one of them has a minimum framerate of 59fps and the other a minimum framerate of 40fps, that's a huge deal, yet almost none of these fuckers actually give you numbers like that.

Average framerates DON'T matter. I want to know how bad it can get. That's what actually matters.
 
ND40oz said:
If you're on VMware, EVC modes are either Intel or AMD, so vMotion won't work at all between platforms. You'll need completely new clusters of all AMD hosts to switch over and you'll need shut down the VMs to do so. It's not exactly just a trivial change of just hardware.
Click to expand...

This is true. Though I must say, surely planned maintenance windows allow for an organization to transition their VMware infrastructure to AMD from Intel.
 
bigbluefe said:
As someone who really only cares about gaming performance, I'm waiting for AMD to release a CPU that can actually beat Intel in single threaded gaming applications. Sure, Intel's a disaster right now, but AMD's been such a disaster they still haven't caught up with Intel's single thread performance, which is pretty pathetic, honestly.

Also, I'm sick of worthless benchmarks. I don't care about average framerates. I want to see minimum framerates.

If the averages between two CPUs are 105fps and 110fps, but one of them has a minimum framerate of 59fps and the other a minimum framerate of 40fps, that's a huge deal, yet almost none of these fuckers actually give you numbers like that.

Average framerates DON'T matter. I want to know how bad it can get. That's what actually matters.
Click to expand...

is wprime that important to you?

what game is single threaded?

cities XL is the "newest" one i can think of
 
N4CR said:
Funny how some enterprise users say AMD migration is a little tricky but they are still doing it or at least giving it a crack, whilst others don't seem to want to do any of that work and act like ostritches.
AMD of 2020 isn't AMD of 2004 anymore and software support is improving across the board. But it still seems to be an emotion-based decision making process for some.
Click to expand...

I work in IT infrastructure. Emotion rarely comes into play, aside from the fact that if you mess up your job may be on the line. The "safe" option tends to be the choice of most in this industry, and right now that means Intel. Things can and do change though, but change tends to be slow. Hardware replacement cycles occur at fixed intervals, and with fixed budgets. It's great that AMD is super competitive in most CPU markets right now, but I can't replace all of my ESX hosts with Epyc-powered blade servers on a whim, I have to wait until the next hardware replacement cycle comes up and I will make the case at that time to do so, if it make sense for our organization.
 
ND40oz said:
It's not the initial amount of work to migrate an instance, it's more work but it's also introducing systems that are completely incompatible with the rest of the environment. I'm just migrating test and dev work loads off of HP DL380 G6s, systems that have been in use for almost 10 years. Everything else in the VMware environment moves down a rung and the new boxes take over the highest tier production workloads. So if I switch over, nothing in my environment can migrate in and out of the highest tier cluster without shut downs. Sorry not happening, I'd need to replace all of my hardware at once, all of my environments to switch over to AMD. It doesn't make any sense to do that, fiscally or level of effort.
Click to expand...

I face the same dilemma. Fortunately for us, our dev environment can be shut down just about any time we need to migrate VMs between hosts.
 
techguymaxc said:
I work in IT infrastructure. Emotion rarely comes into play, aside from the fact that if you mess up your job may be on the line. The "safe" option tends to be the choice of most in this industry, and right now that means Intel. Things can and do change though, but change tends to be slow. Hardware replacement cycles occur at fixed intervals, and with fixed budgets. It's great that AMD is super competitive in most CPU markets right now, but I can't replace all of my ESX hosts with Epyc-powered blade servers on a whim, I have to wait until the next hardware replacement cycle comes up and I will make the case at that time to do so, if it make sense for our organization.
Click to expand...

tell your organization with all the security holes intel has they could be in one of those cool data breach articles.

safe option pfffft!
 
Master_shake_ said:
tell your organization with all the security holes intel has they could be in one of those cool data breach articles.

safe option pfffft!
Click to expand...

The highly-publicized side channel attacks e.g. Spectre/Meltdown have multiple layers of mitigations in software and hardware. No CPU is perfectly secure, neither is any software. My job is to ensure these mitigations are in-place.
 
techguymaxc said:
The highly-publicized side channel attacks e.g. Spectre/Meltdown have multiple layers of mitigations in software and hardware. No CPU is perfectly secure, neither is any software. My job is to ensure these mitigations are in-place.
Click to expand...

no CPU?

then why isn't any AMD cpu on those lists?

could it be your own bias?
 
techguymaxc said:
This is true. Though I must say, surely planned maintenance windows allow for an organization to transition their VMware infrastructure to AMD from Intel.
Click to expand...

Not just to AMD, but also back to Intel, perhaps more than once as business needs dictate.

While I agree that it's obviously well withing the realm of possibility, it also represents a reduction in operational flexibility that may not be acceptable to an enterprise.
 
Master_shake_ said:
no CPU?

then why isn't any AMD cpu on those lists?

could it be your own bias?
Click to expand...

Yes, no CPU is 100% secure.

Here's an AMD-specific CPU vulnerability:
https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2019-9836

Here's AMD's own words about a different vulnerability:

AMD does not believe the PortSmash issue (https://seclists.org/oss-sec/2018/q4/123) is related to previously found speculative execution issues like Spectre. Instead, AMD believes the issues are related to any processor that uses simultaneous multithreading (SMT), including those from AMD, that is vulnerable to software that exposes the activity of one process to another running on the same processor. We believe this issue can be mitigated in software by using side-channel counter measures. For example, OpenSSL, which was used in the researcher’s proof of concept, has already been updated to address this type of attack.
Click to expand...

This isn't a sporting event, and AMD isn't your favorite team. They are a for-profit company, competing with Intel for your money (and mine).
 
IdiotInCharge said:
Prove that those patched vulnerabilities are exploitable in techguymaxc's environment first.
Click to expand...

The SWAPGS instruction is found in most Intel CPUs released after 2012, including those released after Ivy Bridge. On the consumer side, the flaw impacts the third generation of Intel Core processors and beyond, although Bitdefender notes that it also presents a grave threat to enterprise users, as well as those using Intel processors on servers.

In a statement, Gavin Hill, Vice President, Datacenter and Network Security Products at Bitdefender said: “Criminals with knowledge of these attacks would have the power to uncover the most vital, best-protected information of both companies and private individuals around the world, and the corresponding power to steal, blackmail, sabotage and spy.”

As with any chip-level threat, anyone using shared computing platforms are among those at the greatest risk. Those using a cloud computing provider could, for example, see an attacker exploit this vulnerability to access credentials and information within their own user space, such as private encryption keys and passwords.
Click to expand...

using it on a server?

pwned.

https://thenextweb.com/security/201...security-flaw-in-all-modern-intel-processors/
 
You must log in or register to reply here.
Back
Top