Monday, March 28th 2016

More AMD Socket AM4 Technical Details Emerge

More details of AMD's upcoming common socket for both its desktop APUs and high-end CPUs emerged from a recent article by Italian tech-site Bits-n-Chips. To begin with, AM4 will be an µOPGA (pin-grid array), in which the pins will continue to be located on the processor package, and contact points on the socket. The package will be square, and 40 mm in length, making it about as big as a current socket FM2+ package. It will have a pin-count of 1,331 pins, a big increase from the 942 pins of AM3+, and 906 pins of FM2+. AMD could continue to develop LGA sockets for its multi-socket capable Opteron processors based on the "Zen" architecture.

The AM4 platform layout will be functionally closer to that of the FM2+ than the AM3+. Besides the integrated memory controller, the northbridge will be entirely located on the processor die; and so the HyperTransport main system bus will be wired internally. Besides hundreds of electrical pins, the AM4 pin-map will consist of memory I/O, integrated graphics I/O, PCI-Express, and the chipset bus; besides other low-level system I/O interfaces. The memory controller on some of the first AM4 chips, such as "Summit Ridge," will natively support DDR4-2400 MHz, and DDR4-2933 MHz through overclocking.
Sources: 1, 2
Add your own comment

50 Comments on More AMD Socket AM4 Technical Details Emerge

#26
PP Mguire
Am*Umm yes they do. Average Joes will not get them very far -- they need OEMs and the rich server-buying crowds on their side, far more than they need the Average Joes.
Intel's current market slide with Xeons seem to disagree. If AMD can mimic performance or surpass it with a cheaper price they will flood the enterprise market like they used to. The only people who need high memory bandwidth to the CPU are APU owners because the GPU needs it.
Posted on Reply
#27
xenocide
eidairaman1id be more afraid of bending pins on a mobo than a CPU
Way more often the CPU Pins degrade than the Motherboard Pins. Hell, once upon a time I got an Athlon 64 X2 that came with pre-bent Pins (the plastic had flexed when Newegg stored them I assume, and it bent a few of the corner pins). Skylake was an engineering issue that affected a single platform--compare that to OPGA which at this point is a fundamental flaw. There's no technical reason to keep the pins on the CPU, much like how we now integrate many components into the CPU itself, it can be completely done away with and nothing is lost.
Posted on Reply
#28
Jism
alucasaThe problem is the pins. Whether it's on CPU or on mobo, it does not matter. They need to invent a new way that does not involve fragile parts (pins).
Here comes BGA lol.... You cant go wrong with an AMD cpu at all. It's simply 'drop' it into the socket without any force. If you need to use force you are doing it wrong. AMD's can only go one way, and thats by looking towards the PIN-layout of both CPU and socket. Anyone not understanding that shoud'nt be replacing CPU's in the first place.

Quad memory channel provides double the bandwidth of a dual channel setup. And since DDR4 moves alot more data then DDR3 specific, you proberly wont need anything higher then 2400MHz for now.

I'd go for quad channel with as low as possible latency's. AMD always had benefit from memory with extreme tight timings (DDR2 period with 3/3/3/9 for example).
Posted on Reply
#29
Caring1
GoodmanMe think , they should even go back to a further idea... like Intel Slot1 (Pentium2)

But only on a small printed board with only the CPU weld on it then you just slide in the slot just like a video card , no more bending pins on the CPU or Motherboard
It would probably be better for CPU heat dissipation as it would stand up like a video card & probably help to keep the Motherboard cooler also?

Anyhow just a thought...

Mikey
Not a bad thought, adding a Processor to the Graphics card would remove the need for it on the Motherboard, and enthusiasts could add more in SLI or CFX boosting overall performance, of course adding eDramm to the CPU negates the need for Dimm slots also if they can mount an adequate amount.
Posted on Reply
#30
TRWOV
GoodmanMe think , they should even go back to a further idea... like Intel Slot1 (Pentium2)

But only on a small printed board with only the CPU weld on it then you just slide in the slot just like a video card , no more bending pins on the CPU or Motherboard
It would probably be better for CPU heat dissipation as it would stand up like a video card & probably help to keep the Motherboard cooler also?

Anyhow just a thought...

Mikey
Asrock had a 939 board with a slot for AM2 CPUs that were installed in such fashion.
Posted on Reply
#31
FlanK3r
Am*For the love of God AMD...do no mess this launch up. The below quote is already making me nervous:



It needs to support 3200MHz out of the box, no excuses...no more repeats of history please, like the slow-as-hell DDR2/3 mixed on-die controllers of the AM3 days where Sandy Bridge was mopping the floor with the Phenoms with 2-3x the bandwidth/performance using the same memory. And no bloody BIOS flashing to support higher frequencies either...

It's all or nothing for AMD now -- and if this will be another stagnant launch like the Faildozer, I will put my PC building hobby to rest for at least another 5-8 years (or until Intel release something that's at least 3-4x faster than my current CPU, whichever comes first)...which won't be good for either Intel nor AMD. DX12 and especially Vulkan, along with the current gen consoles -- will make it that much easier to hold out for a real upgrade.
Some boards will support higher dividers inside ,-)
Posted on Reply
#32
cdawall
where the hell are my stars
FlanK3rSome boards will support higher dividers inside ,-)
Anyone who didn't assume this is a bit silly ;)
Posted on Reply
#33
jahramika
newtekie1Which is more likely? You drop your CPU, or you mis-handle it and touch one of the pins and bend it. OR You drop something onto your motherboard, in the small time the socket is uncovered and it just happens to hit that small little 1"x1" square target and bends some pins? It doesn't matter that you've never bent a pin on a CPU, I've never bent a pin on a motherboard, the fact is the pins on the CPU make them more prone to damage.



Yes, it very well would have mattered. The thin PCB is what causing the bend. Pins might help a little, but not much. Pins are solidly connected to anything, they float in the socket with a very small force holding them in place. That is why they are so easily pulled out with the heatsink when the TIM has a strong enough bond between the CPU and Heatsink. I've warped CPUs with pins in the past.

Not having support under the middle of the CPU, where the majority of the force from the cooler is focused is what caused skylake to bend.
Posted on Reply
#34
jahramika
Well I dropped my CPU inserting it it was not damaged. Thing is I dropped it on the motherboard CPU socket and yes those pins are unfixable once bent! Thing is when you handle a CPU with pins you see them and are cautious. With pins on the socket you cant really see them until its to late or the sleeve cuff of your shirt just touches those pins and will the cloth catches. I prefer on chip pins myself at least you can straighten a slightly bend one.
Posted on Reply
#35
jahramika
Am*For the love of God AMD...do no mess this launch up. The below quote is already making me nervous:



It needs to support 3200MHz out of the box, no excuses...no more repeats of history please, like the slow-as-hell DDR2/3 mixed on-die controllers of the AM3 days where Sandy Bridge was mopping the floor with the Phenoms with 2-3x the bandwidth/performance using the same memory. And no bloody BIOS flashing to support higher frequencies either...

It's all or nothing for AMD now -- and if this will be another stagnant launch like the Faildozer, I will put my PC building hobby to rest for at least another 5-8 years (or until Intel release something that's at least 3-4x faster than my current CPU, whichever comes first)...which won't be good for either Intel nor AMD. DX12 and especially Vulkan, along with the current gen consoles -- will make it that much easier to hold out for a real upgrade.
Posted on Reply
#36
jahramika
LOL ya really but if they fail someone will buy them and continue with a new management group!
Posted on Reply
#37
TRWOV
*off-topic*

This is what I was talking about in a previous post



You installed that on Asrock 939 boards with "Future CPU port"
Posted on Reply
#38
newtekie1
Semi-Retired Folder
jahramikaThing is I dropped it on the motherboard CPU socket and yes those pins are unfixable once bent!
No, there are plenty of cases of people fixing bent pins on the CPU socket.
Posted on Reply
#39
Thefumigator
JismRead source:
The CPU/NB was already located on the CPU since the Thuban if i'm not mistaken.
Since the very first athlon 64/opteron, which made a huge impact in performance at the time.
Posted on Reply
#40
Thefumigator
TRWOV*off-topic*

This is what I was talking about in a previous post



You installed that on Asrock 939 boards with "Future CPU port"
I had that board, NF3 upgrade, with Nforce 3 chipset and it was 754, I added AM2 later on with the adaptor.
Posted on Reply
#41
scorpion_amd13
Personally, I'm glad they kept the pins on the CPU. LGA sockets are a pain to work with, you can easily snag or bend one of those damn pins and because they're not straight (they're shaped closer to the letter Z than anything else), they are extremely hard to fix. On the other hand, when the pins are on the CPU, they are straight and very easy to bend back into shape using a razor or the blade from a cutter, and it only takes mere seconds. Sure, if they are bent too far out of shape they are likely to break, but then again, LGA socket pins aren't exactly sturdier (quite the opposite, I would say).

And you don't need to worry about storing the damn socket protection cap for the CPU socket, you only have to store the CPU's package. Also, it's really easy to tell which pin is bent too much when the pins are on the CPU, whereas with pins on the socket... it's not as easy (you can't stare down the length of the rows of pins because your view will always be obstructed by something).

I've also seen an incorrectly mounted stock cooler (only 3 out of the 4 push-pins of the retention system properly attached to the motherboard) push the CPU over the pins in the socket and thus bend them. Now, that was back in the days of the Pentium 4, hopefully that is not an issue anymore (can't say I've tested this on newer LGA sockets, for obvious reasons). At any rate, this simply cannot happen when the pins are on the CPU instead of the socket.

Quite frankly, the only two advantages LGA sockets have when compared to classic sockets are the lower profile (well, usually, anyway) and the fact that if you damage one of the pins beyond repair, you brick the motherboard instead of the CPU. But otherwise, LGA sockets are a nightmare by comparison.
Posted on Reply
#42
Frick
Fishfaced Nincompoop
ThefumigatorI had that board, NF3 upgrade, with Nforce 3 chipset and it was 754, I added AM2 later on with the adaptor.
I had one of the best boards of all time, the Dual-Sata2. S939 (so less point with the upgrade board but 754 to AM2 is a two generation jump and is hugely cool); fast AGP, PCIe, decent overclocker, less than €50..
Posted on Reply
#43
AsRock
TPU addict
john_There are cases with bent pins on the motherboard and cases with bent pins on the cpu. In every case you have to be careful. Then there is also the Skylake chips that were bending under force. That wouldn't have happened if the pins were on the cpu.
Better to have bent pins on a mobo as it's normally cheaper to replace, how even if you do have bent pins it's easier to put them right on a cpu than a mobo for sure.
Posted on Reply
#44
john_
AsRockBetter to have bent pins on a mobo as it's normally cheaper to replace, how even if you do have bent pins it's easier to put them right on a cpu than a mobo for sure.
It depends. On Intel platform there are some really expensive boards and I don't think it is a rare case having a $100 motherboard with an $80 cpu. But forgetting for a moment what is cheaper, if you have to replace the motherboard this means plenty of work to remove it and put it back. If your cpu's pins bent, things are much easier. Also a motherboard replacement could lead to an OS not booting and having to install it again and you will also have to verify your OS key. And what happens if your Win 10 key is married to that board?
Posted on Reply
#45
Am*
PP MguireIntel's current market slide with Xeons seem to disagree. If AMD can mimic performance or surpass it with a cheaper price they will flood the enterprise market like they used to. The only people who need high memory bandwidth to the CPU are APU owners because the GPU needs it.
Most enterprise workloads need high memory bandwidth by default. Ones that don't will use higher amounts of lower performing CPUs instead anyway (hyperthreaded quads/hexa cores etc) -- and since AMD want to make AM4 a jack-of-all-trades socket with APUs in the pipeline further proves my point.
Posted on Reply
#46
PP Mguire
Am*Most enterprise workloads need high memory bandwidth by default. Ones that don't will use higher amounts of lower performing CPUs instead anyway (hyperthreaded quads/hexa cores etc) -- and since AMD want to make AM4 a jack-of-all-trades socket with APUs in the pipeline further proves my point.
Zen APUs are quoted to support HBM, while the desktop/server class CPUs support up to 8 channels. Anything that doesn't require GPU graphics workloads on the processor memory pipeline doesn't require a ton of memory bandwidth like I said before. If the APUs can come with HBM on them to support the graphics then no, the platform doesn't require said bandwidth capabilities. The server showdown will come down to core performance. If 8 channels quoted is true, it'll already have an edge on Intel but I bet it'll be quad which has been sufficient the past 4 years.
Posted on Reply
#47
cdawall
where the hell are my stars
The 8 channel stuff is dual 4 channel controllers just like the current server parts have dual dual channel it's another lie amd pushes onto the public there is next to no scenario when they bandwidth is usable except the one benchmark amd runs on a pr slide.
Posted on Reply
#48
AyyMD
newtekie1Damn, I was hoping for LGA. Pins on the processor is so 1990's.
And it's also the reason I'm avoiding AMD CPUs until they have LGA socket types, I will be AyyMD only for GPUs.
Posted on Reply
#49
Totally
FrickI don't think so. Is there a reason for the pins to be on the CPU?
None, there isn't a technical reason for either configuration. iirc Intel only made the move to pass the cost and warranty headaches to motherboard manufacturers, and in my personal opinion, I think that AMD didn't have the clout even less so nowadays to make MB manufacturers deal with the pins.
Posted on Reply
#50
nkata
My first PC had a 286-12.5MHz processor soldered to a daughterboard with a prospected upgrade potential to a 386-16. Needless to say I sold the PC before the upgrade.

I have a liking for pins on the CPU. Installation is usually a breeze. They are getting very thin though which may be an issue with so many.
Posted on Reply
Add your own comment
Nov 21st, 2024 11:58 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts