Lessons Learned: Ryzen Launch

[Atari2600]

Was just mulling it over in my head, what could AMD have done differently with the Ryzen launch?


On one side, with their current FX CPU revenue/profit being in the gutter, they are financially forced to launch the product as soon as possible to raise some money.

Converse to that, releasing the product means reviewers etc get their hands on it and opine online about the CPU, which can form (positive or negative) opinions among potential purchasers - these opinions can be hard to shift.


On the assumption that the time constraint cannot be shifted - and that the CPUs, Mobos and bios we had at launch were to remain unchanged, is there anything AMD could have done to make Ryzen's foibles more palatable?



I've shortlisted a number of ways AMD could have handled the launch, please feel free to add any <VIABLE> additional options (bearing in mind the constraints above) for what they could have done.

Who knows, maybe for Zen2 they may use this thread as a point of reference for how to improve matters.

 

[sm625]

They were way too stingy with engineering samples.

 

[Panino Manino]

"Admit to the public and reviewers that launch CPU/Bios/Motherboards are in Beta stage".

Its so hard to say that first results wouldn't be the best that Ryzen is capable?

 

[dark zero]

"Admit to the public and reviewers that launch CPU/Bios/Motherboards are in Beta stage".

Its so hard to say that first results wouldn't be the best that Ryzen is capable?

It was to surprise Intel and they did big time. Intel couldn't counter that.

 

[LTC8K6]

It looks like if they'd waited a month things would have been considerably smoother...not sure why they couldn't arrange to push the reviews back a month.

 

[LTC8K6]

It was to surprise Intel and they did big time. Intel couldn't counter that.

I think Intel is feeling pretty good about the RyZen launch and performance, actually.

Intel appears to have made the right moves, which was basically no moves.

 

[dark zero]

I think Intel is feeling pretty good about the RyZen launch and performance, actually.

Intel appears to have made the right moves, which was basically no moves.

Actually it has a LOT to moves to cancel Zen once at all. RyZen is a teasing for Raven Ridge which is Zen with a GPU and with no errors from RyZen and that is even worse.

 

[LTC8K6]

Actually it has a LOT to moves to cancel Zen once at all. RyZen is a teasing for Raven Ridge which is Zen with a GPU and with no errors from RyZen and that is even worse.

Intel certainly might react if it seems like they need to.

Right now, there's very little sign that they need to change anything, imo.

Not sure why Raven Ridge would be that much more of a threat than previous APUs were.

Previous APUs already had far better IGPs than Intel, and they still do.

There is also the possible problem of creating an IGP that is as good as, or better, than some of your own video cards.

 

[philosofool]

The original post has bought in to AMD's "Ryzen needs optimization and BIOS tweaks" story. I will believe that when I see it. I think what we've seen of Ryzen is basically the story on Ryzen. We might see 2-3% gains, but AMD didn't waste the spotlight by releasing a CPU that was going to become awesome! in 3 months.

It's a very good processor at a much lower price than the competition. There are a few benchmarks where it is well behind Intel, there are some where it is ahead. It does not have any "bugs" that we know of; that is, the processor does not encounter critical failures in any task. I am seriously considering a Ryzen 5 or Ryzen 3 for my next machine.

AMD does not owe early adopters anything. Every benchmark they published was legit. The fact that some people wanted/expected it to be better is not something for which they should be held accountable. Maybe they're guilty of the standard "hyping the product", but if that's grounds for compensation, we're all owed money by every company that's ever sold us anything.

 

[imported_jjj]

Not much could be done differently without changing the timing. Not that the launch is over, there is still some confusion about the 8 cores and they should have 6 and 4 soon.The launch s not the end, in a way it's a beginning.

The launch needs to be a positive surprise but there was the gaming problem and few can be bothered to test power and efficiency as that was a big plus.
Considering the quality of the reviews, they might as well publish their own numbers for gaming, efficiency, Win 7, memory scaling.
They need to make it easier for folks to choose the DRAM they need.
They could have made a limited Signature series for everything shipped in March, denoting an early adopter and include some stickers and a shirt. While at it, include a Ryzen die keychain in every box- never seen anything like that but it would be cool.
They should find a way to let users OC while keeping a user defined single core boost, that would do wonders for Ryzen's image as folks can settle for a decent MT OC and keep ST.
What can be fixed must be fixed before the Ryzen 5 launch and do that launch ASAP. That's an opportunity to show Ryzen in a better light.
They need to force reviewers to do a better job so publishing their own data for R7 before R5 arrives would also help with that. Show them the diff between high FPS and normal FPS, the difference between peak power and efficiency and more.
AMD should have some mobos with the Ryzen color theme,marketing is what it is but it works. Maybe add some Ryzen branded cable ties, a USB hub
Some laptops with 8 core Ryzen wouldn't hurt either.
They should talk more about sticking with the socket for a few gens about the Infinity fabric.Highlight efficiency to get folks excited about Raven Ridge.

So mostly it's about communication and marketing at this point.

 

[whm1974]

I think things will improve for RyZen once BIOS updates and software patches start rolling out. http://www.phoronix.com is at least is claiming that RyZen does make a great Linux system for users who need good multi-threading performance on a shoestring budget.

 

[alexruiz]

I think most of the situation could have been improved if providing more samples and earlier to the motherboard manufacturers. They are the ones making sure the CPU has a good performing house.

I understand AMD wanted to avoid leaks as much as possible, but when you see the wide differential in performance between reviews that we have traced to using different boards, it was imperative to have at least the top boards performing well, and specially, consistently well.
The boards makers should have been given more samples much sooner.

More samples to the press? Not at all, have them speculate.
The existing samples should have gone to the people doing the real engineering work.

 

[Kenmitch]

Other.

The past was used to predict the future bit the MB makers in the arse! Once they saw the potential of Ryzen it was too late/ lose to launch.

Above omits the rest of my thoughts which get me a infraction/warning most likely.

 

[Malogeek]

1. Work with Microsoft way before the release on anything requiring optimizing regarding thread allocation and CCX usage
2. Work with motherboard vendors way before the release on BIOS microcode and memory support
3. Release Ryzen with a running start and dominate the market (requires 1 & 2)

 

[guachi]

The big failure was with motherboard manufacturers. Gotta have a motherboard to use with the chip.

However, if the R5 is competitive (and I think it will be) all the motherboard supply, BIOS, Windows scheduler problems should be gone.

It's a chance to turn a good play into a great one.

 

[formulav8]

more samples and earlier to the motherboard manufacturers

The biggest problems for mobo makers was because AMD completely redesigned the firmware 3 months ago.

 

[Sable]

tbh, I think they were thinking the processor itself would to the talking. And to a certain extent it has, it's the most interested I've been in a new CPU since I got my 2500k. I know I still don't NEED an upgrade but they've created an itch.

The gaming performance looks absolutely fine to me, the only thing holding me back is my generally cautious nature and the odd thing I've read about RAM speeds and BIOS niggles. This is where people are a bit O_O. Iron them out, give my brain time to think on a new build and I'll be right in there.

 

[scannall]

tbh, I think they were thinking the processor itself would to the talking. And to a certain extent it has, it's the most interested I've been in a new CPU since I got my 2500k. I know I still don't NEED an upgrade but they've created an itch.

The gaming performance looks absolutely fine to me, the only thing holding me back is my generally cautious nature and the odd thing I've read about RAM speeds and BIOS niggles. This is where people are a bit O_O. Iron them out, give my brain time to think on a new build and I'll be right in there.

I'll build a Ryzen system this summer. I've never been big on being an unpaid beta tester. ;-) It's a great CPU, and I look forward to others working the kinks out first.

 

[DrMrLordX]

1. Work with Microsoft way before the release on anything requiring optimizing regarding thread allocation and CCX usage
2. Work with motherboard vendors way before the release on BIOS microcode and memory support
3. Release Ryzen with a running start and dominate the market (requires 1 & 2)

You're assuming that MS and/or the mobo OEMs were willing to commit much time/effort to the launch. Look at AMD's market share prior to the Ryzen launch. Would you take them seriously?

Ryzen sold like hotcakes. Nobody (outside AMD) expected that at the top. Nobody.

 

[lolfail9001]

You're assuming that MS and/or the mobo OEMs were willing to commit much time/effort to the launch. Look at AMD's market share prior to the Ryzen launch. Would you take them seriously?

Well, the story right now is that AMD did not leave them any time to speak of anyways.

 

[AMDisTheBEST]

I picked other. Do a surprise launch is what I would suggest, meaning disclose no info on their new products to the point that we don't even know ryzen exist. Then suddenly out of the blue sky BOOM, new ryzen CPU. Then they do the hype, hype it as high as they can.

 

[Rifter]

What they did was done by choice, and i dont think they should have done it any different.

They made the choice to keep it tight with no leaks, you cannot give ES to board makers any earlier than they did or it WILL get leaked, thats a guarantee. They wanted to blindside intel, and they succeed in doing just that.

They had to blindside intel really, and far enough before SL-X to get some sales in before intel launches SL-X at reduced prices to counter zen. Intels current HEDT looks very overpriced and has horrible price/perf vs zen and AMD took advantage of that. This im sure will change with SL-X so AMD had to cash in now.

Really all i think they could have done was get some software fixes done first, no reason they could not have worked with MS to get the windows scheduler fixed pre launch. And started to work with game companies sooner.

 

[Kenmitch]

Well, the story right now is that AMD did not leave them any time to speak of anyways.

Story from who?

 

[Malogeek]

You're assuming that MS and/or the mobo OEMs were willing to commit much time/effort to the launch. Look at AMD's market share prior to the Ryzen launch. Would you take them seriously?

You mean the only other maker of x86 processors capable of running Windows? Yes I do expect MS to take them seriously

 

[lolfail9001]

Story from who?

http://www.legitreviews.com/one-motherboard-maker-explains-why-amd-am4-boards-are-missing_192470