AMD will launch AM4 platform in March 2016 says industry source

[Percy1983]

To be honest is it matches the intel mid/high (ie i5) on performance and price they have me.

Marginal cases will also go AMD for me, they need the money more.

[DanceswithUnix]

Intel could have released an 8-core years ago if they'd wanted to (infact they did, SB-E is an 8-core with 2 cores disabled), but they didn't because there was no need to.

Intel disabled two cores to get the thermals right. ISTR you could buy the part with all 8 cores as a Xeon, but the clock speeds sucked for gaming etc use so that didn't get sold as a consumer part. So no, they couldn't release a consumer 8 core part, or at least not a good one.

And no, I haven't forgotten BD. I waited a bit, and bought PD, which I am still using and very happy with. In fact I am unlikely to buy Zen straight away for the same reasons I haven't upgraded to the latest i7, my 8350 is still fast enough for my needs.

[scaryjim]

I'm also a little wary of AMD chipsets / drivers ...

... the good old days of random nForce-related BSODs ...

Blaming AMD for NVidia's dodgy chipsets seems a little rich.

I've never had a problem with an AMD chipset. Third party chipsets for AMD, yes. AMD's own chipsets, never.

[Sumanji]

You don't recall Intel's 6 series SATA bug? Or the Haswell USB bug? Or the TSX extensions they had to retroactively withdraw from processors after they discovered a bug in its implementation? The latter of which is of limited impact to consumers, though.

Sure, there have been some bugs - which Intel have been fairly forthright and open about (for a tech company anyway).

But historically, given the choice between an Intel chipset vs. AMD / nV / Via / SIS etc., I know which I'd pick... and I'm definitely a self-confessed AMD fan! Massive caveat is that I haven't used any recent AMD chipsets, so certainly my opinion needs reevaluating

Blaming AMD for NVidia's dodgy chipsets seems a little rich.

A platform is only as good as its weakest link - if AMD had to rely on nV to deliver performance chipsets for their CPUs, yes I can and will blame them!

It's good to know that you have not had any problems with recent AMD chipsets. I look forward to hopefully having the same quality of experience as you, but that trust and reputation needs to be earned, not assumed!

[badass]

All this talk of Zen might be a little misguided. I believe AMD were planning an excavator based 28nm APU to be released on AM4. Bristol ridge IIRC. It's probably that you'll see early ish 2016.

[DanceswithUnix]

All this talk of Zen might be a little misguided. I believe AMD were planning an excavator based 28nm APU to be released on AM4. Bristol ridge IIRC. It's probably that you'll see early ish 2016.

That's fine. AMD need to have motherboards out there. There is currently just one AM3+ uATX motherboard worth having (the Asrock 970 chipset one) and I presume FX cpus are selling out as prices and availability seem to be getting worse recently. FM2+ motherboards are OK if you don't want high end performance.

Just a shame that it looks like Bristol Ridge won't have a 6 core option as it will still be on 28nm and AMD are still pushing the graphics part hard. Still, BR now and possible Zen upgrade later is fine.

[watercooled]

Sure, there have been some bugs - which Intel have been fairly forthright and open about (for a tech company anyway).

But historically, given the choice between an Intel chipset vs. AMD / nV / Via / SIS etc., I know which I'd pick... and I'm definitely a self-confessed AMD fan! Massive caveat is that I haven't used any recent AMD chipsets, so certainly my opinion needs reevaluating

My point is that I don't understand why you think AMD chipsets need to earn some reputation compared to Intel? It's not the case that Intel 'have always been rock solid' either, even relatively speaking.

A platform is only as good as its weakest link - if AMD had to rely on nV to deliver performance chipsets for their CPUs, yes I can and will blame them!

But that wasn't the case at all? Nvidia were an alternate chipset supplier for AMD just like they were for Intel. Up until Intel blocked them from making compatible chipsets a while back, that is. There was a time Nvidia forced you to have an Nvidia chipset if you wanted to use SLI on both AMD an Intel platforms. But of course they had to concede that artificial requirement after the aforementioned occurrence - the alternative being no SLI on Intel platforms.

It's good to know that you have not had any problems with recent AMD chipsets. I look forward to hopefully having the same quality of experience as you, but that trust and reputation needs to be earned, not assumed!

Again, considering AMD have been in the chipset business for a very long time, why would any reputation have to be assumed?

I'm not meaning to be argumentative at all, but I'm not sure I understand where you're coming from.

[shaithis]

My point is that I don't understand why you think AMD chipsets need to earn some reputation compared to Intel? It's not the case that Intel 'have always been rock solid' either, even relatively speaking.

Sheesh....I think I have to go back to the boards where you had to add cache modules before I can remember an intel chipset with issues....430FX?

But that wasn't the case at all? Nvidia were an alternate chipset supplier for AMD just like they were for Intel. Up until Intel blocked them from making compatible chipsets a while back

And that was probably one of the best moves Intel made. I had the "pleasure" of dealing with multiple nVidia motherboards.....their stability was reminiscent of pc chips motherboards with via or SIS chipsets........

Again, considering AMD have been in the chipset business for a very long time, why would any reputation have to be assumed?

Because Intels chipsets have been solid for a lot longer. Intel have the reputation, AMD need to build a similar rep. The state of the AM3+ market is what many have been looking at since early 2012....the boards didn't exactly set the world on fire, although you could argue that board manufacturers were to blame, not AMD.....but AMD had a long time to do something about it....yet didn't

[watercooled]

I provided multiple examples of defects with Intel chipsets in an earlier posts? Personally I don't consider them to have been solid because of issues like that. I wouldn't avoid them, but I would consider it illogical to call them more solid or reputable than AMD with that in mind.

[Percy1983]

Personally I have no problem with AMD chipsets, got an old Pentium 4 machine at my father in laws which was/still is fine.

Setup a friends PCwith a 785G chipset, still rock solid.

I am running 2 PCs (A88X and A75) and you guessed it both are perfect, never even thought anything about it until mentioned here.

[christreleaven]

Well I was going to build a Skylake machine, using the i7 6700k, but now I'm tempted to wait for the new AMD processors.

I'll have to have a good think about this one.

[bert7]

Zen will be good for what you are going to pay, it may not quite perform as per intel, but then processors are so good now days it wont matter to most people, they will be fast and cheaper than intel, chuck it on a board with HBM2 and bingo VR, 4k, whatever, it will do it.

[MercutioUK]

It seems there's a good number of folks either too young or forgetful of AMD around the P4 clarksdale era. They were the first up with 64bit processors and worlds ahead of Intel after spending around a decade doing "me too" chips to go in Intel sockets.
I'm not for a second saying this will be the winner they need but they've totally blown Intel out of the water before when they seemed unassailable so it could happen again IF they get the magic right.

[MRFS]

I've been an Intel fan for many years. Nevertheless, the DMI 3.0 link
on their latest chipsets is forcing a low ceiling on the max bandwidth
of high-performance solid-state storage. I think AMD would do well
to adopt of policy of embracing the NVMe standard, and at a minimum
begin to design AMD CPUs and matching chipsets that support a full
x16 lanes to NVMe RAID controllers with x16 edge connectors. The main
design reason is the symmetry that results from four U.2 ports @ x4 PCIe
lanes (4 @ x4 = x16), and the advantages that come with support for
all modern RAID levels. Remember also that PCIe 4.0's clock will
oscillate at 16 GHz, and upcoming non-volatile DRAM solutions promise
to require much higher upstream bandwidths.

[MRFS]

Case in point:
http://www.thessdreview.com/daily-ne...t-performance/

See this Comment by Eric Johnson (right on point):

"The DMI 3.0 connection between the processor and z170 chipset is limited to about 4GBps, that could be a bottleneck and the reason you got similar numbers for the 2-device and 3-device arrays. Since the m.2 4x PCIe 3.0 interface can also do about 4GBps, I'm thinking just one drive will be enough to saturate that bottleneck in the not-too-distant future."

PCIe 3.0 uses a 128b/130b "jumbo frame" (130 bits / 16 bytes)

x1 PCIe 3.0 lane = 8 GHz / 8.125 bits per byte = ~ 1.0 GB/second

DMI 3.0 = x4 PCIe 3.0 lanes = ~ 4.0 GB/second

One NVMe SSD also uses x4 PCIe 3.0 lanes i.e. same bandwidth
as the DMI 3.0 link.

We would like to RAID four NVMe SSDs and do so with
adequate upstream bandwidth e.g. equivalent to x16 PCIe 3.0 lanes.

[DanceswithUnix]

Supposedly the dual core replacement for Carrizo-L will go on this socket as well: http://wccftech.com/amd-bristol-ston...e-apus-listed/

That would be really nice, from the equivalent of socket AM1 up to Zen on a single platform.