Intel Announces Sunny Cove Brings 18% IPC Improvement

True and although I think AMD is currently more secure, they were also less prone to focus since they've less market. This might change, I hope they can still be as secure as this would give them a great argument but right now... it's worth mention but nothing more. We shall see in 1 or 2 years if AMD can keep their momentum.

AMD's current offerings are more secure as far as anyone knows. However, it may be one of those "security through obscurity" things. More effort and research has been done regarding attacks for Intel processor based systems due to their prevalence in the server market.
 
AMD's current offerings are more secure as far as anyone knows. However, it may be one of those "security through obscurity" things. More effort and research has been done regarding attacks for Intel processor based systems due to their prevalence in the server market.

At this point, it's hard to see any out of order CPU as being secure except through obscurity. Which, if you've worked in security before, you know that security through obscurity isn't.

Especially if you're a target as large as Intel.
 
Yeah that's what I meant, security through obscurity and we shall see if that will be as secure as Apple thought it was in the days where they made their no AV needed claim lol.
 
if this is true i might buy intel just one more time - single core performance is still important to me
 
I really hope AMD puts a lot of money in testing this being that they plan (if we can believe the youtube rumors) on going 4 wide SMT in Zen4.
 
Until someone can prove the exploit on a AMD processor then yeah it's safer then Intel. Fair or not that is reality.

Safer at the moment? Yes. Enough for them to build a whole marketing brand/segment around? No, not really, and that is the point. It's a nice footnote to go along with their exceptional value at the moment, and nothing more.
 
Until someone can prove the exploit on a AMD processor then yeah it's safer then Intel. Fair or not that is reality.

A new security flaw for EPYC Zen was communicated to AMD in February and fixed yesterday. EPYC is so low volume that researchers don't spend the same amount of time looking for vulnerabilities as they do for Intel Xeons.
 
AMD's current offerings are more secure as far as anyone knows. However, it may be one of those "security through obscurity" things. More effort and research has been done regarding attacks for Intel processor based systems due to their prevalence in the server market.


And so you are inferring they could not simultaneously deal with these issues? How hard is it to have a division set up to test AMD? Or is bone idle society to attribute for this
 
And so you are inferring they could not simultaneously deal with these issues? How hard is it to have a division set up to test AMD? Or is bone idle society to attribute for this

It means exactly what it means. Globally (not just within one company), more man hours are spent hacking Intel CPUs than AMD CPUs, so flaws are more likely to be discovered on Intel. Just like more man hours are spent hacking Windows than Linux and Macs. Could they spend equal time hacking Intel and AMD? Yes. Do they? No, because Intel has the much larger marketshare.
 
It means exactly what it means. Globally (not just within one company), more man hours are spent hacking Intel CPUs than AMD CPUs, so flaws are more likely to be discovered on Intel. Just like more man hours are spent hacking Windows than Linux and Macs. Could they spend equal time hacking Intel and AMD? Yes. Do they? No, because Intel has the much larger marketshare.

Brand specific hacking is a weak argument, hackers don't break system encryption based on brands it is a very generic practice and to assume it doesn't require the same level of intrusion is just completely hearsay and has no probative value. If we assume mainstream has shown intrusion and AMD has half the mainstream share then I am calling this one out.

Maybe AMD is actually just better and why does that even worry anyone. I just like that there is a underdog out there capable of giving that incompetent bastion of capitalism a bloody nose and you would think someone at Intel would have pulled the finger out by now.

As to testing intrusions both are x86 so it is not really that hard and the fact nothing has come up is probably because some sparky at AMD thought about that and is probably now wanted by Intel
 
Maybe AMD is actually just better and why does that even worry anyone.

They're newer- that's about it. Skylake is getting very old, it's older than some of the Bulldozer spins AMD was peddling for so long. Skylake alone is probably the most owned x86 CPU architecture ever. It's not hard to see that it's the primary target for attackers and researchers.

At best, what's going to be found in AMD CPUs is going to be different, but the principle remains: it's impossible to create software code without bugs, and modern hardware is just software fixed in purpose and translated to wires and transistors on a die.

What matters more to enterprise users is not the prevalence of bugs, but the support structure available to correct them, and here Intel wins by an obvious landslide.

If AMD continues to gain marketshare and as their products age in the market, they'll also have to invest significantly into their support side as they'll be targeted more and more.
 
They're newer- that's about it. Skylake is getting very old, it's older than some of the Bulldozer spins AMD was peddling for so long. Skylake alone is probably the most owned x86 CPU architecture ever. It's not hard to see that it's the primary target for attackers and researchers.

At best, what's going to be found in AMD CPUs is going to be different, but the principle remains: it's impossible to create software code without bugs, and modern hardware is just software fixed in purpose and translated to wires and transistors on a die.

What matters more to enterprise users is not the prevalence of bugs, but the support structure available to correct them, and here Intel wins by an obvious landslide.

If AMD continues to gain marketshare and as their products age in the market, they'll also have to invest significantly into their support side as they'll be targeted more and more.


Skylake was 2017 barely 2 years ago which is not old and given how the Core I uArch has been around since 2008 odd it is long in the tooth and with Intels huge workforce maybe somebody should have thought about this in the security level of the CPU.

AMD is gaining share and partner, AMD outlay will be directly proportional to the investing partner and a few big ones are starting to come out with support for AMD's progressive marketing. If marketing and position is disputed then I think Dr Lisa Su knows just how to do that well, she was able to polish a turd into revenue with bulldozer and ARM cores, she is very capable of positioning AMD into good waters going forward and unlike Hector Ruiz no amount of Intel money will buy her complicit obedience in putting AMD into idle mode, under her AMD is hell bent on growing its market and it is going to bug Intel make no mistake there.
 
Skylake was 2017 barely 2 years ago

no.

the Core I uArch

Core isn't an architecture, it's a brand.

If marketing and position is disputed then I think Dr Lisa Su knows just how to do that well, she was able to polish a turd into revenue with bulldozer and ARM cores, she is very capable of positioning AMD into good waters going forward and unlike Hector Ruiz no amount of Intel money will buy her complicit obedience in putting AMD into idle mode, under her AMD is hell bent on growing its market and it is going to bug Intel make no mistake there.

This unceasing hero worship is getting old.
 
no.



Core isn't an architecture, it's a brand.



This unceasing hero worship is getting old.

It is not hero worship, but she took AMD from poor investor confidence into a investor positive public listed company, success for a PLC is based on investor confidence, and clearly it is working and I am assuming they are people that actually know what they are doing.
 
A new security flaw for EPYC Zen was communicated to AMD in February and fixed yesterday. EPYC is so low volume that researchers don't spend the same amount of time looking for vulnerabilities as they do for Intel Xeons.

Yeah thats what a normal issue looks like. A issue was found, duplicated and then fixed and a new firmware was released then disclosed to the public. While Intel has had to tell people to disable Hyper-threading and leave it that way cause they have no way to fix the issue. Which one looks worse to you? Also I think Google is looking quite closely since they are using Epyc platforms and low and behold that is the group that reported the issue.
 
Yeah thats what a normal issue looks like. A issue was found, duplicated and then fixed and a new firmware was released then disclosed to the public. While Intel has had to tell people to disable Hyper-threading and leave it that way cause they have no way to fix the issue. Which one looks worse to you? Also I think Google is looking quite closely since they are using Epyc platforms and low and behold that is the group that reported the issue.
Well Intel doesn't have to worry anymore. It is happening so often that people are not surprised anymore.
I hope Intel and AMD keep trading blows. Makes it good for us. :)
 
Back on topic, so it seems like Raspberry Pi's are all you need to hack these days.

Far, far less. What you need is the social engineering skill and opportunity to get access. A Pi is just cheap and easy and commonplace enough to not easily give away the identity of the attacker.

And that's not any closer to the topic than your hero worship posts.
 
Brand specific hacking is a weak argument, hackers don't break system encryption based on brands it is a very generic practice and to assume it doesn't require the same level of intrusion is just completely hearsay and has no probative value. If we assume mainstream has shown intrusion and AMD has half the mainstream share then I am calling this one out.

Maybe AMD is actually just better and why does that even worry anyone. I just like that there is a underdog out there capable of giving that incompetent bastion of capitalism a bloody nose and you would think someone at Intel would have pulled the finger out by now.

As to testing intrusions both are x86 so it is not really that hard and the fact nothing has come up is probably because some sparky at AMD thought about that and is probably now wanted by Intel

What the hell are you even talking about? Intel and AMD utilize completely different architectures, their flaws are not going to be the same. Sure, there are some similarities, but it doesn't mean flaws are universal or even accessed the same way. What is happening now is the attacks focus on exploiting Intel archtecture, then seeing if it works on AMD for the longshot that there is enough similarity that it does. It's like saying two games can be exploited the same way because they use the same Unreal engine. It does not work that way.
 
Brand specific hacking is a weak argument, hackers don't break system encryption based on brands it is a very generic practice and to assume it doesn't require the same level of intrusion is just completely hearsay and has no probative value. If we assume mainstream has shown intrusion and AMD has half the mainstream share then I am calling this one out.

Maybe AMD is actually just better and why does that even worry anyone. I just like that there is a underdog out there capable of giving that incompetent bastion of capitalism a bloody nose and you would think someone at Intel would have pulled the finger out by now.

As to testing intrusions both are x86 so it is not really that hard and the fact nothing has come up is probably because some sparky at AMD thought about that and is probably now wanted by Intel

They do target what is most likely to pay off or what is likely to have the largest payoff though.

Maybe AMD is actually better, but to rely on that for your security when there is really no way to know is foolishness. They're only stating that fact.


Also, while they may both accept x86 instructions streams for execution, what happens inside them are VERY different.
 
It's like saying two games can be exploited the same way because they use the same Unreal engine. It does not work that way.

Well, I actually think your analogy is good: both are x86 CPUs that at a high level work very similarly, as necessitated by the need to be able to run the same code. The basis for many of the attacks pertain to how these CPUs work, while the details will necessarily differ enough that the attack code will vary significantly.

AMD had more time to refine their architecture to defend against hypothetical attacks brought up by researchers with Zen, while Intel had Skylake taped out far longer and with a vast majority of the marketshare, have borne the brunt.
 
Yeah thats what a normal issue looks like. A issue was found, duplicated and then fixed and a new firmware was released then disclosed to the public. While Intel has had to tell people to disable Hyper-threading and leave it that way cause they have no way to fix the issue.

In reality Intel has said the contrary. Also you are comparing apples to oranges here. Whereas a simple encryption mistake in the PSP must be easy to fix, speculative threats as Spectre are much more difficult to mitigate/fix.
 
In reality Intel has said the contrary. Also you are comparing apples to oranges here. Whereas a simple encryption mistake in the PSP must be easy to fix, speculative threats as Spectre are much more difficult to mitigate/fix.

What Intel recommends and what is actually secure is very different with the latest vulnerability Zombieland. As for apple and oranges it's not my fault you picked a weak bug to complain about on the AMD side. If Intel keeps up this pace they will have to change their logo to a piece of Swiss Cheese as they have many large vulnerabilities that have forced them to eat quite a bit of performance to fix. it's one of the main reasons my city awarded a contract for Epyc servers because governments are paranoid about security and I expect that is where AMD will gain quite a bit of traction.
 
What Intel recommends and what is actually secure is very different with the latest vulnerability Zombieland. As for apple and oranges it's not my fault you picked a weak bug to complain about on the AMD side. If Intel keeps up this pace they will have to change their logo to a piece of Swiss Cheese as they have many large vulnerabilities that have forced them to eat quite a bit of performance to fix. it's one of the main reasons my city awarded a contract for Epyc servers because governments are paranoid about security and I expect that is where AMD will gain quite a bit of traction.

You talked about Intel says about HT, but Intel says the contrary of what you pretend. I mentioned the last known bug for EPYC, just to highlight how a bug has been present in EPYC since launch and only detected four months ago. This example shows that researchers aren't paying the same attention to AMD due to its minuscule datacenter marketshare. Goverments often make idiotic things.
 
You talked about Intel says about HT, but Intel says the contrary of what you pretend. I mentioned the last known bug for EPYC, just to highlight how a bug has been present in EPYC since launch and only detected four months ago. This example shows that researchers aren't paying the same attention to AMD due to its minuscule datacenter marketshare. Goverments often make idiotic things.

Oh I am not pretending, what they say in public vs closed doors are two different things, they will openly admit behind closed doors under certain uses your better off to disable Hyper-threading. Even with that they still offer no guaranty you will be totally protected. You keep pretending AMD is just as vulnerable yet you cant prove it, while it's no issue at all to see all the articles and news about ZombieLoad or MDS any lots of other exploits.
 
It is not hero worship, but she took AMD from poor investor confidence into a investor positive public listed company, success for a PLC is based on investor confidence, and clearly it is working and I am assuming they are people that actually know what they are doing.

And just to clarify, you're not wrong about AMD being the better and safer buy in most cases at the moment. What you are wrong about is why that is the case.
 
And just to clarify, you're not wrong about AMD being the better and safer buy in most cases at the moment. What you are wrong about is why that is the case.

I would assume performance is a large reason
 
They do target what is most likely to pay off or what is likely to have the largest payoff though.

Maybe AMD is actually better, but to rely on that for your security when there is really no way to know is foolishness. They're only stating that fact.


Also, while they may both accept x86 instructions streams for execution, what happens inside them are VERY different.

Again this is hearsay, if I put you in a witness stand would you testify to fact or opinion?

The generalisation or discrimination based argument that Intel users are richer or more Intel equates to more chance of scoring is absurd, this is like saying R Kelly sexually exploited woman on the basis of allegations without it being tested.

You cannot confirm researchers haven't extensively tested AMD

You cannot confirm who attacks will be on and their hardware

This is hearsay or opinion and fact. Until said researchers confirm otherwise it appears Intel is more vulnerable, which supports the increased number of reported intrusions.
 
Again this is hearsay, if I put you in a witness stand would you testify to fact or opinion?

The generalisation or discrimination based argument that Intel users are richer or more Intel equates to more chance of scoring is absurd, this is like saying R Kelly sexually exploited woman on the basis of allegations without it being tested.

You cannot confirm researchers haven't extensively tested AMD

You cannot confirm who attacks will be on and their hardware

This is hearsay or opinion and fact. Until said researchers confirm otherwise it appears Intel is more vulnerable, which supports the increased number of reported intrusions.

Well you have to admit that there's more to win accusing R Kelly than accusing me (average joe)... but yeah both should be "tested / trial" before any conclusion is drawn.
Also even if a researcher says he checked and they're not vulnerable does it really mean they're not ? As I see it right now, AMD is more secure but the next few months will tell us if it holds. (Hopefully yes).
 
Oh I am not pretending, what they say in public vs closed doors are two different things, they will openly admit behind closed doors under certain uses your better off to disable Hyper-threading. Even with that they still offer no guaranty you will be totally protected. You keep pretending AMD is just as vulnerable yet you cant prove it, while it's no issue at all to see all the articles and news about ZombieLoad or MDS any lots of other exploits.

They say in public that disabling HT doesn't solve the security issue. Also I didn't say that AMD is just so vulnerable. I said your claim AMD is more secure is a false security cvlaim because researchers aren't paying the same attention to analyze EPYC vulnerabilities, I gave a example of a security flaw that remained unnoticed since EPYC launch until February of this year, and explained why researchers pay less attention to AMD: "EPYC is so low volume that researchers don't spend the same amount of time looking for vulnerabilities as they do for Intel Xeons."
 
They say in public that disabling HT doesn't solve the security issue. Also I didn't say that AMD is just so vulnerable. I said your claim AMD is more secure is a false security cvlaim because researchers aren't paying the same attention to analyze EPYC vulnerabilities, I gave a example of a security flaw that remained unnoticed since EPYC launch until February of this year, and explained why researchers pay less attention to AMD: "EPYC is so low volume that researchers don't spend the same amount of time looking for vulnerabilities as they do for Intel Xeons."

Well technically that is one way to acheive security............
Apple uses this logic...
 
I would assume performance is a large reason

Performance/$$$. Intel still holds the absolute performance crown at the moment. And the current exploits/security issues plays very little into affecting purchasing decisions in the grand scheme of things, for the moment.

Again this is hearsay, if I put you in a witness stand would you testify to fact or opinion?

The generalisation or discrimination based argument that Intel users are richer or more Intel equates to more chance of scoring is absurd, this is like saying R Kelly sexually exploited woman on the basis of allegations without it being tested.

You cannot confirm researchers haven't extensively tested AMD

You cannot confirm who attacks will be on and their hardware

This is hearsay or opinion and fact. Until said researchers confirm otherwise it appears Intel is more vulnerable, which supports the increased number of reported intrusions.

How is it absurd? You pick the biggest target because it has the highest chance of payback. Simple logic, which you seem to have a very hard time comprehending.

I love this place, the trigger is real and the fan bros are loose

Pot meet kettle...
 
Performance/$$$. Intel still holds the absolute performance crown at the moment. And the current exploits/security issues plays very little into affecting purchasing decisions in the grand scheme of things, for the moment.



How is it absurd? You pick the biggest target because it has the highest chance of payback. Simple logic, which you seem to have a very hard time comprehending.



Pot meet kettle...
Intel has the performance per dollar crown? Clock to clock, how much of a difference are we talking?
 
Intel has the performance per dollar crown? Clock to clock, how much of a difference are we talking?

Yeah no they do not, they have the perf crown. Perf / dollar is AMD. Well unless I missed a boat and 2 post-its ?
 
Intel has the performance per dollar crown? Clock to clock, how much of a difference are we talking?
Yeah no they do not, they have the perf crown. Perf / dollar is AMD. Well unless I missed a boat and 2 post-its ?

In general, they don't have the $ / performance crown, primarily because they excel at stuff that is dependent on single-core performance. Of course, in general, your desktop users have no use for more than four cores, and gamers have no use for more than six- so having six for general desktop stuff and eight for gamers is the right kind of overkill.

For most people, whatever's cheapest and has four real cores is going to provide the maximum performance they can see. For gamers, six cores and clockspeed.

For enthusiasts that do all kinds of other stuff, the answer is that it depends, but mostly, AMD.
 
In general, they don't have the $ / performance crown, primarily because they excel at stuff that is dependent on single-core performance. Of course, in general, your desktop users have no use for more than four cores, and gamers have no use for more than six- so having six for general desktop stuff and eight for gamers is the right kind of overkill.

For most people, whatever's cheapest and has four real cores is going to provide the maximum performance they can see. For gamers, six cores and clockspeed.

For enthusiasts that do all kinds of other stuff, the answer is that it depends, but mostly, AMD.
So single core speed on Intel is 50% or more? How much?

No games use the extra cores for running something else while they game? Streamers? Ventrilo? There are MANY reason even gamers could use more than 4/6.
I would say 8 since it's not much more. Single core? If the extra bit is worth it for competitions, then yes.
 
Back
Top