It's official. Eight-core Ryzen 7 chips will start at just $329.

[mrgreenthump]

AMD and AIBs said the microcodes were released 3 weeks before launch. They could be trying to cover the rear, no one knows.

Microcode change has no real effect on hardware though.. And if it did, we would've seen a push of the launch.

[Artorius]

At least those issues were only on these server/workstation chips so far and all the bugs with the platform will hopefully be ironed out by the time the mainstream chips arrive.

AMD hasn't released any of their server SKUs yet, the 1800X, 1700X and 1700 are all supposed to be mainstream consumer parts.

We do have a demo for Naples though:



And some "AMD's own HEDT" rumours that are probably just rumours but you never know. It seems extremely overkill.

- - - Updated - - -

3 weeks is nothing in software development. I can imagine that it is even less in hardware development.

For something that was designed over more than 6 years, yeah... Literally nothing.

[Lillpapps]

For something that was designed over more than 6 years, yeah... Literally nothing.

This doesn't matter, it will still need changes. It still takes a set amount of time to polish, tweak and QA and 3 weeks is not a long time.

[Artorius]

Seriously, you don't have these issues, at least not as bad, when intel launches a new architecture. The board manufacturers get it right when dealing with intel. The difference here is the time AMD gave them to have things ready. That's all there is to it.

X99 launch was also ridiculously buggy. And Skylake still has a problem with prime numbers to this day, AMD isn't the only one with launch problems.

- - - Updated - - -

This doesn't matter, it will still need changes. It still takes a set amount of time to polish, tweak and QA and 3 weeks is not a long time.

You missed the post entirely. I was agreeing with you. It was designed over 6 years, of course they wouldn't be done with fine tuning in 3 weeks.

[Gray_Matter]

They rushed this out and did not give anyone else, including but not limited to, motherboard manufacturers, Microsoft's people who handle how things are scheduled or software devs.

I don't think you can include MS and the devs in that. I think they didn't give the MB guys enough time but anyone who has dealt with MS will realize that they do things at their own pace. Also, devs will only start changing their code when they get an idea of the size of the market. Or if you pay them. In this case, those changes could easily make a 10% difference in speed but they won't be able to work on that until they get machines (including the motherboards) in front of them and some guides from AMD.

I do think that the MB guys share some of the blame. If they are only updating bios, etc, that is. Getting last minute changes that impacts your bios shouldn't impact the number of boards you make. I get the feeling that they were being very cautious with their volumes.

- - - Updated - - -

You missed the post entirely. I was agreeing with you. It was designed over 6 years, of course they wouldn't be done with fine tuning in 3 weeks.

You don't know IT very well. There is always a mad rush at the end, no matter how much time there is

[Artorius]

You don't know IT very well. There is always a mad rush at the end, no matter how much time there is

This isn't IT. It's engineering. I'm in the electrical engineering field so I guess I know it well enough =)

Thing is, Ryzen was originally planed to launch on 2015 and it obviously didn't. If AMD knew they'd do it only in 2016 from the beginning, it wouldn't even be released on 14LPP, they would use 10LPE instead. They had to rewrite the entire firmware 4 months ago and that's why their partners, which obviously couldn't keep up with the pace, ended up with unfinished products. There are a lot of things that were validated ages ago, like the 13.33x multiplier. And some of them which are still fairly new and immature, like the 16x multiplier.

AMD could've postponed the launch for another month, to ship it with better support and give more time for people to test and come up with shit that actually makes sense. Right now the platform is getting better each day with newer BIOSes but this isn't exactly something that should happen after the product is released. The problem with postponing it for another month is that they end up losing money due to the time the product stays in the market, and if they were to wait for everything to be fixed they'd only launch this thing next year. There are some significant differences between the last batch of engineering samples and the actual product in the shelves, and when you have to deal with changes so often it isn't as simple to have the entire platform ready for the launch.

X99 had bugs, Skylake has bug, all launches have problems. The problem here is that the existing problems are interfering with performance in a significant way and this might also hurt them because people testing this usually don't really know why they're getting the results they are, they just care about putting those results in the internet telling people what they should buy.

I don't think AMD did the wrong thing releasing it in this current state. They need a new product and Ryzen is amazing. I'm sure the alarming problems will all be fixed soon enough, putting the product in the market with problems also accelerates bugfixing because they can't leave it bugged there. In development you can drag it for longer.

[moremana]

Here is some hope. MS may be releasing a patch on the Tuesday 14th to address the Ryzen scheduler.

Source

[Fascinate]

I doubt it will come on the 14th, it would be awesome tho as my mobo/cpu get delivered the 15th lol.