All AMD Ryzen CPUs to be Unlocked

Burticus said:
Well sure. Let people overclock them like crazy, then when they melt/explode/vanish into a wormhole they have to buy another one.
Click to expand...
I'm guessing sarcasm? I might be new around here (2006) but I could have sworn their was an OC somewhere in the site name... [H]mmmmm...
 
Zarathustra[H] said:
ECC by its design necessarily increases latency. Unless you absolutely need ECC it is thus usually better to go without, from a performance perspective.
Click to expand...
I think ECC is only ever unnecessary in a pure media consumption box like a HTPC.
Memory errors can lead to silent data corruption. If one values the data that is handled by a computer, one uses ECC.

ECC also gives some protection against the Rowhammer attack. Recently, it was demonstrated how Rowhammer + memory deduplication could be used by an attacker in one virtual machine to attack another one. So if you run untrusted code in a VM, ECC is also a necessity.

https://fahrplan.events.ccc.de/congress/2016/Fahrplan/events/8022.html
https://media.ccc.de/v/33c3-8022-memory_deduplication_the_curse_that_keeps_on_giving

Zarathustra[H] said:
I think that as we start using more and more RAM in our systems thkgun, the risk of a flipped bit in RAM is going to increase, and maybe ECC will need to become a standard feature.
Click to expand...
I don't expect that ECC will become a standard feature without some kind of regulatory intervention. Non-ECC is cheaper, plus most people are unaware of how severe and frequent memory errors can be.
 
Regulatory intervention? From the same clowns that want to legislate back doors into encryption, pc power usage because computers are clearly using way more power hungry now than in the P4 days, won't do anything about our broken patent system, punish responsible disclosure of security issues, relate bandwidth to trucks and tubes?

No, people who have no clue and are willfully ignorant should not be creating regulations on anything.
 
The "clowns" who regulate the environmental and safety aspects of a computer are not at all the same who want to put backdoors in our encryption or refuse to fix the patent system.
 
mjz_5 said:
Where would you see a benefit by using ECC? Have you ever lost something because of a memory error?
Click to expand...

Lots of people have servers with ZFS, where ECC is very highly reccomended. Intel limited ecc support on their quad core chips to Xeons. I'd like to see ECC supported on AM4 like it was with the AM3+ chips.
 
I am thinking that Kyle is sitting in that new chair he is demoing playing on a Ryzen computer from AMD and just laughing at all of us over the little tid bits he keeps throwing out at us and the way we are gobbling it all up.
 
dcds1 said:
I am thinking that Kyle is sitting in that new chair he is demoing playing on a Ryzen computer from AMD and just laughing at all of us over the little tid bits he keeps throwing out at us and the way we are gobbling it all up.
Click to expand...

But does that matter in the end you don't have to buy Ryzen if it does not fit your needs. AMD can't make money just showing benchmarks in the end it has to be a product that sells and with some smart marketing people sell well.

But what I heard is that AMD sent press samples last week (or this week). In the end I hope that the product is good and the motherboards are great so there is a reason to buy Ryzen.
 
dcds1 said:
I am thinking that Kyle is sitting in that new chair he is demoing playing on a Ryzen computer from AMD and just laughing at all of us over the little tid bits he keeps throwing out at us and the way we are gobbling it all up.
Click to expand...
I do not have a Ryzen CPU sample.
 
DeathFromBelow said:
Lots of people have servers with ZFS, where ECC is very highly reccomended. Intel limited ecc support on their quad core chips to Xeons. I'd like to see ECC supported on AM4 like it was with the AM3+ chips.
Click to expand...

I will just leave this here...

You can expect 1 bit flip per GB per week, roughly on average.

Some references:
http://www.scientificamerican.com/article/solar-storms-fast-facts/
"Extensive background radiation studies by IBM in the 1990s suggest that computers typically experience about one cosmic-ray-induced error per 256 megabytes of RAM per month."

http://lambda-diode.com/opinion/ecc-memory
" So the probability of having at least one bit error in 4 gigabytes of memory at sea level on planet Earth in 72 hours is over 95% .

Note that you don't need 72 hours of continuous operation for that to happen. If you leave your computer on from 8 in the morning to midnight, that'll mean that you'll have a bit error in 4.5 days with probability exceeding 95% . Modern operating systems use all the available memory as a disk cache. Very easy to fill it up."

This study raises the probability even higher.
http://www.cs.toronto.edu/~bianca/papers/sigmetrics09.pdf

Based on my own operating experience, these probabilities are in the ball-park.
 
lolfail9001 said:
And for whatever reason that worries me.

If it is not of particular secret, how far ahead in advance did you have 6700k sample and other earlier Intel review samples?
Click to expand...
Usually no more than a week, sometimes a few days.
 
As a guy who was an enthusiast back in the AMD Athlon Barton 2500+ days, I hope AMD will Ryzen again. Ugh.......Sorry.
 
I am not expecting to Ryzen be any good compared to Intel counterpart. You are better with 7700K for $340, i bet it will perform faster than Ryzen. Ryzen does come in 8/16 and i believe it will cost somewhere between 500-600 and that's a lot compared to 7700K. Also it is a decent cpu against 2011-3 socket lineup however here are some concerns i have with Ryzen.
1. Very low number of PCIe compared to Intel 40 PCIe -> i can have three cards running in 16/16/8 setup.
2. Rather weak chipset feature wise
3. Only dual memory channel.
 
Luka said:
I am not expecting to Ryzen be any good compared to Intel counterpart. You are better with 7700K for $340, i bet it will perform faster than Ryzen. Ryzen does come in 8/16 and i believe it will cost somewhere between 500-600 and that's a lot compared to 7700K. Also it is a decent cpu against 2011-3 socket lineup however here are some concerns i have with Ryzen.
1. Very low number of PCIe compared to Intel 40 PCIe -> i can have three cards running in 16/16/8 setup.
2. Rather weak chipset feature wise
3. Only dual memory channel.
Click to expand...

1. Tri fire or Tri sli is pretty much worthless and a waste, heck even two cards these days are becoming a thing of the past. Only 4k people look for dual card setups for the most part.

2. What chipset feature do you think it needs? Seems to have most things that are needed.

3. Quad channel memory is waste unless your at server level, at consumer level it only affects a benchmark by a couple percent and nothing in most anything else.
 
Luka said:
I am not expecting to Ryzen be any good compared to Intel counterpart. You are better with 7700K for $340, i bet it will perform faster than Ryzen. Ryzen does come in 8/16 and i believe it will cost somewhere between 500-600 and that's a lot compared to 7700K. Also it is a decent cpu against 2011-3 socket lineup however here are some concerns i have with Ryzen.
1. Very low number of PCIe compared to Intel 40 PCIe -> i can have three cards running in 16/16/8 setup.
2. Rather weak chipset feature wise
3. Only dual memory channel.
Click to expand...
And how did the 7700K compare to the 2600K again? Not sure why you would compare a 4 core to an 8 and even remotely expect the 8 core to win out. AMD will have 4 cores for that.
 
Well what happened to the quite dated claim these were going to sport 32 cores? I remember it was a big deal before we even had a glimpse of IPC and clocks.
 
Shintai said:
Some server SKUs with 4 dies in MCM on a package will have 32 cores.
Click to expand...

Thanks. Okay, server. I know they aren't meant to be overclocked, but ... will those platforms offer overclocking?

I keep dreaming of building a huge overclocked hog to serve family members via some sort of silent thin clients.
 
JustReason said:
And how did the 7700K compare to the 2600K again? Not sure why you would compare a 4 core to an 8 and even remotely expect the 8 core to win out. AMD will have 4 cores for that.
Click to expand...

I think Kaby Lake is going bring better gaming performance than any version of Ryzen CPU for less money and that's where buyer will go for.
 
AMD currently validates Ryzen 6C / 12T @ 3.3 GHz base but NOT 8C / 8T. The goal is clearly to recycle the failed 8C dies.

 
cageymaru said:
AMD currently validates Ryzen 6C / 12T @ 3.3 GHz base but NOT 8C / 8T. The goal is clearly to recycle the failed 8C dies.

Click to expand...

I'm sure they have some sort of plan this time around with what to do with the chip that fail validation, maybe they will release a bunch of them in certain price brackets.
Some people suggested that they would have batches where they can't fix the SMT/mOP cache bug and dump them as 8C8T.
 
MotionBlur said:
Which makes sense - they've done that for years since the Phenom days IIRC.
Click to expand...

I guess some of the all knowing pundits can start wiping the egg off their faces. I know one guy who is the star on Tom's Hardware and less so on SemiAccurate, Juanrga, who must be splitting his gut. He was wrongly predicting top base clock speed of 3.0 GHZ for Ryzen before New Horizons and he was one of the first to jump on the no 6 core Ryzen movement. There 2 sources were WCCFTech and some other obscure Chinese site . And every lesser web site just jumped on their bandwagon to hell. Give kudos to Kyle for not jumping into that fray of incredulity. I do NOT accept their pricing scenario for the 8 cores either. Why would the Chinese source for pricing and skus be right on that when he was wrong on 6 cores??? I could be wrong, but I stick to my gut instincts and not the spineless who jump on bandwagons.
 
os2wiz said:
I guess some of the all knowing pundits can start wiping the egg off their faces. I know one guy who is the star on Tom's Hardware and less so on SemiAccurate, Juanrga, who must be splitting his gut. He was wrongly predicting top base clock speed of 3.0 GHZ for Ryzen before New Horizons and he was one of the first to jump on the no 6 core Ryzen movement. There 2 sources were WCCFTech and some other obscure Chinese site . And every lesser web site just jumped on their bandwagon to hell. Give kudos to Kyle for not jumping into that fray of incredulity. I do NOT accept their pricing scenario for the 8 cores either. Why would the Chinese source for pricing and skus be right on that when he was wrong on 6 cores??? I could be wrong, but I stick to my gut instincts and not the spineless who jump on bandwagons.
Click to expand...

I think that's the right approach. We all want Zen to succeed and be competitive regardless if you're an Intel or AMD fan boy, but I think the Polaris launch reminded me (and others on here) of being patient and ignoring all of the crazy rumors (good or bad) until we're either holding it in our hands doing our own testing or hopefully before then Kyle doing an in-depth test.

It should be an exciting next couple of weeks as official information comes out.
 
MotionBlur said:
I think that's the right approach. We all want Zen to succeed and be competitive regardless if you're an Intel or AMD fan boy, but I think the Polaris launch reminded me (and others on here) of being patient and ignoring all of the crazy rumors (good or bad) until we're either holding it in our hands doing our own testing or hopefully before then Kyle doing an in-depth test.

It should be an exciting next couple of weeks as official information comes out.
Click to expand...

You see though, Video cards are of a whole different class of hardware. When it comes to the Ryzen CPU's we are guaranteed that they will be considerably faster than the existing FX series of processors. I could not care less if it beats Intel's latest or not, but it will be faster than what they are selling now. (That said, my FX builds are pretty fast too but, the chipset is getting old and does not support any of the newest stuff like DDR4.)
 
Kyle I bet you have a full review already written, probably have the chips in your house/office, and just can't say squat until NDA lifts.

Let us know if they are looking sweet so far? You can let us know without breaking NDA muahahah
 
tangoseal said:
Kyle I bet you have a full review already written, probably have the chips in your house/office, and just can't say squat until NDA lifts.

Let us know if they are looking sweet so far? You can let us know without breaking NDA muahahah
Click to expand...
Nope. Have not seen hide nor hair of anything Ryzen. I wish I had, but I think it is a ways off yet.
 
Kyle, if we assume that Ryzen launches in early March, when would you expect chips to arrive at [H]?
 
chithanh said:
Kyle, if we assume that Ryzen launches in early March, when would you expect chips to arrive at [H]?
Click to expand...
I have seen CPUs arrive a month early, I have seen CPUs arrive two days before launch, and I have CPUs not arrive at all. So I have no expectations. AMD is still not talking to me for the most part.
 
Well I hope you get one Kyle tho I think you wont have but a few days with it for a review. Still look forward to your review and see what you think of it.
 
You must log in or register to reply here.
Back
Top