[RUMOUR] AMD's Zen CPU rumoured to compete with high end intel for $300

[cozz]

Tweaktown did a live stream to their Facebook with more information about it and why they posted the news 

[dalekphalm]

4 minutes ago, cozz said:

Tweaktown did a live stream to their Facebook with more information about it and why they posted the news 

Care to link something about that?

 

Anyway, this is straight up bullshit. As others have said, it either competes with "high end Intel", or it's $300. If it performed like a 6900K, I could see them maybe undercutting Intel by 20% max (Probably not even that much). No way in hell it'll be less than $800 - and even that would be pretty shockingly low.

 

I can certainly believe it competes with high end Intel. It's just not gonna cost $300 while doing so. And anyone who believes it is straight up naive.

[MageTank]

2 minutes ago, cozz said:

Tweaktown did a live stream to their Facebook with more information about it and why they posted the news 

I can sum that up without a livestream. "It gets clicks". 

 

Posting a reddit rumor as a source is about as low as it gets on the speculation train. Especially when it is completely outside of the realm of possibilities. If AMD prices their 4c, 8t SKU at $250, they already step on Intel's i7 and i5 lineups, even more so if it's unlocked. If they price their 8c, 16t SKU at $500, it's literally half the price of the 8c, 16t Intel versions. Even if it's clock speeds are say, 20-30% lower, for half the price, people that need a lifting CPU on a budget will flock to this. They already flock to those cheap Xeon ES chips on Ebay for that very reason. 

 

I've said this before, and I'll repeat myself for the thousandth time. Zen does NOT need to match Skylake (Or Kaby, since again, they both have the same IPC). If it matches Haswell, and bring's DDR4, it will be enough. Haswell in terms of IPC, is only 5% slower than Skylake. DDR4 does a lot of the heavy lifting in those benchmarks when comparing Skylake to Haswell. The biggest hurdle AMD has, is the fact that 14nm is new to them, and they need to make sure their chips can OC decently compared to Intel. They need to win the hearts of the enthusiasts again. This doesn't require them to beat Intel, but offering very similar performance at a very competitive price will be more than enough. 

[DrMikeNZ]

Quote

BIOS or UEFI is actually built into the CPU, so only AMD can update the "BIOS" or microcode. All overclocking must take place within the Operating system

 

Right now it takes up to 30 minutes to clear the BIOS. If you remove the CPU and place it on another motherboard, it'll have the same settings applied as on the previous board. So debugging is a nightmare

WTF. AMD would have to pay me $300 to use one of those processors.

 

I am under the impression the speculated price was for their 4C/8T processor, and either there was some troll regurgitating click-bait, or they didn't know the difference.

[dalekphalm]

4 minutes ago, DrMikeNZ said:

WTF. AMD would have to pay me $300 to use one of those processors.

 

I am under the impression the speculated price was for their 4C/8T processor, and either there was some troll regurgitating click-bait, or they didn't know the difference.

I wouldn't put much stock in the on chip bios or the "30 minutes to clear BIOS" or any of that. I've heard no confirmation of any such things.

[Energycore]

I really don't trust tweaktown lol

[patrickjp93]

1 hour ago, MageTank said:

Almost as dead as Kaby's IPC boost.

 

#MageTankvPatrick2016Part8Episode2RevengeOfTheMyth

Sorry but you can deny GB4 all you want, and you're still wrong.

[patrickjp93]

#IntelForPOTUS2016

[Lays]

1 hour ago, MageTank said:

Almost as dead as Kaby's IPC boost.

 

#MageTankvPatrick2016Part8Episode2RevengeOfTheMyth

 

3 minutes ago, patrickjp93 said:

Sorry but you can deny GB4 all you want, and you're still wrong.

wat

[MageTank]

4 minutes ago, patrickjp93 said:

Sorry but you can deny GB4 all you want, and you're still wrong.

I still don't even know what GB4 is. Impossible to deny what I do not understand. However, every single other source on the net disagrees with you, and your only source is outdated (and uses the exact same testing methodology of the dozens of other results I gave you). We've been over this before in regards to Iris Pro 580, Keller, etc. You also know that unlike you, I admit when I am proven wrong. If you have a source (again, your geekbench results were shown to be vastly wrong compared to every other result on that site) I'll gladly take a look. Otherwise, bask in the glory of being on the receiving end of my terrible jokes. Not everybody gets this special attention after all.

[DrMikeNZ]

36 minutes ago, huilun02 said:

*unofficial posts rumor about Zen*
"Omg AMD is setting itself up for disappointment for making all this hype"
"Zen will blow everyone away with performance just look at the (lack of) benchmarks"
"Zen will be a failure just look at the (lack of) benchmarks"

 

Wtf is wrong with this community?

This is where things went wrong:

36 minutes ago, huilun02 said:

*unofficial posts rumor*

[Space Reptile]

1 hour ago, Lays said:

wat

fanboy and someone who likes to trash-talk AMD in every thread , just ignore him 

[Mooshi]

AMD lost me generations ago. Iwant so badly for them to succeed, but it redic at this point to be wowed by rumors. 

[MageTank]

1 minute ago, Space Reptile said:

fanboy and someone who likes to trash-talk AMD in every thread , just ignore him 

He knows who Patrick is. He is just confused due to the lack of context of what Patrick and I are talking about. Patrick lives in a world where Kaby has up to a 6% increase in IPC. His only source are his "friends" and their geekbench scores. Problem is, their 7700k's are the only 7700k's scoring that high. Every single other 7700k scores lower, on par with that of 6700k's at the same clock speeds. I even clocked my 6700k and ran the test just to show him. He also thinks he's "won" every argument that he has ever had with me, but anyone that has ever been on this forum the past 2 years knows that to be false. 

 

It's why I like Patrick to be honest. He is an entertaining fellow. 

[Space Reptile]

1 minute ago, MageTank said:

Patrick lives in a world where Kaby has up to a 6% increase in IPC.

even if that where true it would be still shit 

2 minutes ago, MageTank said:

their 7700k's are the only 7700k's scoring that high.

"my dad works at nintendo"

3 minutes ago, MageTank said:

He also thinks he's "won" every argument that he has ever had with me, but anyone that has ever been on this forum the past 2 years knows that to be false. 

i know that type , they put up one counter argument or some random number w/o any context and ignore that you instantly disprove em or that they have been false the entire time , had the honors w/ that kind in that Hubble thread , some will never learn others refuse to 

 

5 minutes ago, MageTank said:

It's why I like Patrick to be honest. He is an entertaining fellow. 

that is true , but sometimes its more frustrating that funny , like an Adam Sandler movie 

[Slyhawk]

when it's out we will know if it's good and worth the price till then stop getting baited and hyped just wait till it's released. 

[IAEInferno]

4 hours ago, TheRandomness said:

You're like a person out of the WCCFTech comments section. You appear on every controversial thread and slam your viewpoint into everyone's face. 

He is not the only one that does that, and not only on WCCFTech.

[patrickjp93]

2 hours ago, MageTank said:

I still don't even know what GB4 is. Impossible to deny what I do not understand. However, every single other source on the net disagrees with you, and your only source is outdated (and uses the exact same testing methodology of the dozens of other results I gave you). We've been over this before in regards to Iris Pro 580, Keller, etc. You also know that unlike you, I admit when I am proven wrong. If you have a source (again, your geekbench results were shown to be vastly wrong compared to every other result on that site) I'll gladly take a look. Otherwise, bask in the glory of being on the receiving end of my terrible jokes. Not everybody gets this special attention after all.

GeekBench 4...

 

No, they were shown to be different, not wrong.

[patrickjp93]

1 hour ago, Space Reptile said:

even if that where true it would be still shit 

"my dad works at nintendo"

i know that type , they put up one counter argument or some random number w/o any context and ignore that you instantly disprove em or that they have been false the entire time , had the honors w/ that kind in that Hubble thread , some will never learn others refuse to 

 

that is true , but sometimes its more frustrating that funny , like an Adam Sandler movie 

Define still shit. Not only is that 1.05^5 = 1.27628 or 27.628% higher IPC than Sandy Bridge (5% per generation on average), but the clocks are also 20% higher. Then there's the fact you get DDR4, AVX2, FMA3, a much more powerful iGPU media engine, and many other capabilities that older chips just plain lack.

 

And if you're about to come back with the core count argument, I've proven here on LTT that core count doesn't necessarily mean crap. With AVX accelerating a common game engine workload, I got 10X the performance while still using just 1 core, vs. the < 4X performance I would have gotten by multithreading scalar code. Amdahl's Law dictates the next performance revolution is vectorization, even as old as the capability is (Jaguar, Bulldozer, and Sandy Bridge for AVX).

[patrickjp93]

1 hour ago, Space Reptile said:

fanboy and someone who likes to trash-talk AMD in every thread , just ignore him 

I don't trash-talk AMD. I merely explain why their current solutions are practically worthless to myself and many other people.

[MageTank]

12 minutes ago, patrickjp93 said:

GeekBench 4...

 

No, they were shown to be different, not wrong.

Yes, I know exactly which benchmark it was. In fact, I provided the list, did I not? 

 

 

On 11/2/2016 at 6:33 AM, MageTank said:

Well, you better intend to do so, because using your exact same Geekbench results, I can show you a dozens of results scoring much lower than that, much more consistent with what Skylake scores. In fact, I'll just go ahead and do that.

 

https://browser.primatelabs.com/v4/cpu/917704

https://browser.primatelabs.com/v4/cpu/917402

https://browser.primatelabs.com/v4/cpu/915287

https://browser.primatelabs.com/v4/cpu/915251

https://browser.primatelabs.com/v4/cpu/915193

https://browser.primatelabs.com/v4/cpu/874959

https://browser.primatelabs.com/v4/cpu/874884

https://browser.primatelabs.com/v4/cpu/858013

https://browser.primatelabs.com/v4/cpu/857993

https://browser.primatelabs.com/v4/cpu/857949

https://browser.primatelabs.com/v4/cpu/857912

https://browser.primatelabs.com/v4/cpu/857848

https://browser.primatelabs.com/v4/cpu/857762

https://browser.primatelabs.com/v4/cpu/857549

https://browser.primatelabs.com/v4/cpu/857202

https://browser.primatelabs.com/v4/cpu/857139

https://browser.primatelabs.com/v4/cpu/856608

https://browser.primatelabs.com/v4/cpu/856552

https://browser.primatelabs.com/v4/cpu/856537

https://browser.primatelabs.com/v4/cpu/726300

https://browser.primatelabs.com/v4/cpu/710789

https://browser.primatelabs.com/v4/cpu/636643

https://browser.primatelabs.com/v4/cpu/636614

https://browser.primatelabs.com/v4/cpu/634429

https://browser.primatelabs.com/v4/cpu/633525

https://browser.primatelabs.com/v4/cpu/633164

https://browser.primatelabs.com/v4/cpu/633138

 

The only "consistent" thing about your Geekbench results, are that they score no higher than the average Skylake does, at the same clock speeds. Don't worry Patrick, you already know I don't shy away from your attempts to "beat me". After all, you are the only person on this forum that even entertains me anymore.

Tell me here and now. Do you think the results from that single Gigabyte crew, are the only accurate results, and that every single one of the results quoted above in my previous post, are wrong? If so, what evidence do you have to support those claims? At 4.5ghz, my results were as followed: 

Let's compare that result, to your favorite result to cite: https://browser.primatelabs.com/v4/cpu/583064

Their 7700k scored 6131 SC, while my 6700k scored 6088 SC. Difference of roughly 0.7%. Their 7700k scored 20243 MC, my 6700k scored 20424 MC. Difference of roughly 0.9%. Both well within margin of error, wouldn't you agree? We also do not have the information of their memory clock speeds, nor can we extrapolate that information from the test itself, given how inaccurate it is when it comes to reporting the proper memory bandwidth. God knows it's measurement of my bandwidth is far off compared to what I score in every other benchmark I use. You yourself repeatedly gave me misinformation as to what kit they were using (as the brands at the specific speed/cas you mentioned simply did not exist). You already know I archive this information. You also know I enjoy doing this to you. Why do you insist on trying when you know it ends exactly the same, every single time? Give me a proper source Patrick, or simply let it be. 

[patrickjp93]

1 minute ago, MageTank said:

Yes, I know exactly which benchmark it was. In fact, I provided the list, did I not? 

 

 

Tell me here and now. Do you think the results from that single Gigabyte crew, are the only accurate results, and that every single one of the results quoted above in my previous post, are wrong? If so, what evidence do you have to support those claims? At 4.5ghz, my results were as followed: 

Let's compare that result, to your favorite result to cite: https://browser.primatelabs.com/v4/cpu/583064

Their 7700k scored 6131 SC, while my 6700k scored 6088 SC. Difference of roughly 0.7%. Their 7700k scored 20243 MC, my 6700k scored 20424 MC. Difference of 

roughly 0.9%. Both well within margin of error, wouldn't you agree? We also do not have the information of their memory clock speeds, nor can we extrapolate that information from the test itself, given how inaccurate it is when it comes to reporting the proper memory bandwidth. God knows it's measurement of my bandwidth is far off compared to what I score in every other benchmark I use. You yourself repeatedly gave me misinformation as to what kit they were using (as the brands at the specific speed/cas you mentioned simply did not exist). You already know I archive this information. You also know I enjoy doing this to you. Why do you insist on trying when you know it ends exactly the same, every single time? Give me a proper source Patrick, or simply let it be. 

No, not within margin of error. Gigabyte's posted score is a 12-run average. Margin of error is roughly 0.2%.

[MageTank]

Just now, patrickjp93 said:

No, not within margin of error. Gigabyte's posted score is a 12-run average. Margin of error is roughly 0.2%.

You clearly do not know how geekbench works. When you run it, it posts the results automatically. You can't run it multiple times, and manually submit an average. Go run the benchmark before using it as your one and only source. Maybe then you would understand these things a little more. 

[patrickjp93]

1 minute ago, MageTank said:

You clearly do not know how geekbench works. When you run it, it posts the results automatically. You can't run it multiple times, and manually submit an average. Go run the benchmark before using it as your one and only source. Maybe then you would understand these things a little more. 

Unless you just don't connect to the internet. Primatelabs allows for submission of validated results external to the benchmark client itself.

[MageTank]

Just now, patrickjp93 said:

Unless you just don't connect to the internet. Primatelabs allows for submission of validated results external to the benchmark client itself.

So now you are making an assumption of their testing methodology, with no way to prove it, and are throwing out a random number that in and of itself, cannot be confirmed? Why do you insist on using this strategy against me every single time we speak? Your entire defense is "You can't prove me wrong", which is sadly true because I lack the ability to travel to this fairy tale realm to obtain the same information you have. 

 

Let's humor your logic for a second though. That they ran it 12 times before submitting a score. Why did they submit two scores? 

https://browser.primatelabs.com/v4/cpu/583275

https://browser.primatelabs.com/v4/cpu/583064

 

24 minute difference between the submission of these two scores. Perfectly capable of running the test 12 times in 24 minutes, I won't argue that, but what purpose does it serve? They are basically the exact same scores, with nothing changed. If they had the information gathered from 12 offline runs, why even bother posting online again, after running another 12 runs, if the score is so similar?