Wednesday, April 5th 2017

AMD Ryzen 3 1200 Specifications Surface

Following its launch of the Ryzen 5 series performance-segment six-core and quad-core processors later this month, AMD could launch entry-level quad-core chips based on the 14 nm "Summit Ridge" silicon in the second half of 2017. This lineup will be called the Ryzen 3 series, and will occupy several sub-$150 price points.

The Ryzen 3 series parts will compete with Intel's Core i3 dual-core "Kaby Lake" processors, and will offer four cores, even if lacking SMT (that's 4 cores, 4 threads), and up to 8 MB of L3 cache, making for a compelling deal against Core i3 "Kaby Lake" dual-core parts that have 2 cores and 4 threads enabled through HyperThreading, and just 3-4 MB of L3 cache. What's more, the Ryzen 3 series chips will come with unlocked base-clock multipliers. One of the prominent Ryzen 3 series SKUs revealed by leaky taps among the motherboard industry is the Ryzen 3 1200.
The Ryzen 3 1200 is a quad-core socket AM4 processor that lacks SMT (simultaneous multi-threading), so you get four cores and four threads. We know from older reports, that Ryzen 3 series quad-core chips feature 8 MB of L3 cache, and have their TDP rated at 65W. The Ryzen 3 1200 comes with clock speeds of 3.10 GHz, and an unknown (likely 3.40 GHz) TurboCore frequency. It could be prices way below the $150 mark, given that better endowed Ryzen 5 quad-core chips with SMT start at $169. Source: ComputerBase.de
Add your own comment

25 Comments on AMD Ryzen 3 1200 Specifications Surface

#1
bogami
Well here's a not much new. Now you can turn off the BIOS 4 cores on R 7 1700X and R 7 1800X ,clock down, and found what to expect. It is a different architecture from intel models. And I am very interested if the limit currently 4 Gh can get beaten. Will R 3 1200 model even be unlocked? Missing a graphical part ? And in this case, too expensive.
Posted on Reply
#2
TheLaughingMan
bogami said:
Well here's a not much new. Now you can turn off the BIOS 4 cores on R 7 1700X and R 7 1800X ,clock down, and found what to expect. It is a different architecture from intel models. And I am very interested if the limit currently 4 Gh can get beaten. Will R 3 1200 model even be unlocked? Missing a graphical part ? And in this case, too expensive.
It is unlocked. There is no graphics part. Not sure about the overclocking as current standings would suggestion the 4 GHz wall is an architecture limit so I doubt it. And the price is a bit high considering its competition in that price range is $150 to $180. They should shoot for about $160 imho.
Posted on Reply
#3
Basard
Maybe we'll ba able to unlock SMT or extra cores on them like the ole glory days.... either way it should be decent if priced right.
Seems like kind of a good idea to wait a year or two and see what spawns from 1st gen Ryzen.... but doesn't it always?
Posted on Reply
#4
ADHDGAMING
Basard said:
Maybe we'll ba able to unlock SMT or extra cores on them like the ole glory days.... either way it should be decent if priced right.
Seems like kind of a good idea to wait a year or two and see what spawns from 1st gen Ryzen.... but doesn't it always?
Its always a good idea to wait .. but i aint gonna xD
Posted on Reply
#5
RejZoR
This is also one CCX, meaning no issues others experience with inter-CCX communications. I think these will eat Core i3 in everything but single threaded tasks that insist on ridiculous clocks...
Posted on Reply
#6
ixi
It's funny in one way.

R7 1700 65w 8 core, R5 models 65w 6 and 4 cores and then R3 dual core with 65w, c'mon :/.
Posted on Reply
#7
danbert2000
ixi said:
It's funny in one way.

R7 1700 65w 8 core, R5 models 65w 6 and 4 cores and then R3 dual core with 65w, c'mon :/.
The TDP ratings from AMD are incredibly coarse. That 1700 pulls more like 90 watts, but I guess AMD wanted to tout that their processor can live in a 65 W TDP without burning up or throttling too hard. I have a 65 W TDP i7 and it continually uses way more power than that whenever it uses the boost clocks. If this is just one CCX, there's no way it uses the same amount as a 1700 at 100% sustained utilization.
Posted on Reply
#8
TheLaughingMan
ixi said:
It's funny in one way.

R7 1700 65w 8 core, R5 models 65w 6 and 4 cores and then R3 dual core with 65w, c'mon :/.
R7 1700 at stock is a much lower clock speed than its bigger brothers. R5 4 cores and R3 4 cores is the same chip and one just has SMT. All of those at 65W seems fine for heat generated. If anything the R5 6 core is the one out of place.
Posted on Reply
#9
horsemama1956
Was thinking of a 5 but might grab one of these until Ryzen 2.
Posted on Reply
#10
Jism
ixi said:
It's funny in one way.

R7 1700 65w 8 core, R5 models 65w 6 and 4 cores and then R3 dual core with 65w, c'mon :/.
The more leaky chips consume or need more power to operate.
Posted on Reply
#11
cdawall
where the hell are my stars
ixi said:
It's funny in one way.

R7 1700 65w 8 core, R5 models 65w 6 and 4 cores and then R3 dual core with 65w, c'mon :/.
You know that isn't a rating for power consumption right?
Posted on Reply
#12
Caring1
RejZoR said:
This is also one CCX, meaning no issues others experience with inter-CCX communications. I think these will eat Core i3 in everything but single threaded tasks that insist on ridiculous clocks...
I was thinking it will compete more with the Pentium line of chips, as the Ryzen has only 2Mb cache and no hyperthreading ...
Posted on Reply
#13
TheLaughingMan
Caring1 said:
I was thinking it will compete more with the Pentium line of chips, as the Ryzen has only 2Mb cache and no hyperthreading ...
I am pretty sure the 2 MB cache is wrong. There is no need to gimp the cache at that point especially when there is a Ryzen 1100 series as well that is going to compete with the Pentium lineup and is expected to be around the $120 mark. The 1200 will likely have at the full 8 MB L3 cache and the 1100 make have less at 2 MB or 4 MB.
Posted on Reply
#14
RejZoR
The charts that I got few days ago all have 8MB cache, even 1100 models. I mean, it's not like you can chop L3 cache in smaller pieces within existing single CCX unit...
Posted on Reply
#15
Totally
Man, I missed the days when there were only two chips per market segments and they were just budget, mainstream, and enthusiast. Already at 9 chips with more on the way.
Posted on Reply
#16
TheLaughingMan
Totally said:
Man, I missed the days when there were only two chips per market segments and they were just budget, mainstream, and enthusiast. Already at 9 chips with more on the way.
Well the market expanded, needs diversified, and pricing segments got smaller but increased in number. Its the shotgun approach. Release something for everyone and see what sells. Adjust your production according to what ending up being your demand areas.
Posted on Reply
#17
cdawall
where the hell are my stars
TheLaughingMan said:
Well the market expanded, needs diversified, and pricing segments got smaller but increased in number. Its the shotgun approach. Release something for everyone and see what sells. Adjust your production according to what ending up being your demand areas.
I am trying to think of how long ago there were two chips per market segment. It was prior to the Athlon k7/P2 so it has been what 20+ years?
Posted on Reply
#18
Totally
cdawall said:
I am trying to think of how long ago there were two chips per market segment. It was prior to the Athlon k7/P2 so it has been what 20+ years?
not quite k7/p2 lasted until 2006 when Conroe arrived, so 10+ years
Posted on Reply
#19
ixi
cdawall said:
You know that isn't a rating for power consumption right?
You will know it when it's released, nor it's estimated as tdp.
Posted on Reply
#20
Frick
Fishfaced Nincompoop
Totally said:
Man, I missed the days when there were only two chips per market segments and they were just budget, mainstream, and enthusiast. Already at 9 chips with more on the way.
I'm not sure what you mean actually. There were CPU's released in every speed avaliable, in 33-200Mhz increments. There has always been a million different Celerons and Athlons out, just to use one example. I count 40 Pentium III models across two sockets. If anything Ryzen is pretty straightforward so far: Four lines (8C/16T, 6C/12T, 4C/8T, 4C/4T) with two or three chips per tier.

What do you mean with P2 btw? To that either means 80286 or Pentium II, neither makes sense in the context.
Posted on Reply
#21
Disparia
Frick said:
I'm not sure what you mean actually. There were CPU's released in every speed avaliable, in 33-200Mhz increments. There has always been a million different Celerons and Athlons out, just to use one example. I count 40 Pentium III models across two sockets. If anything Ryzen is pretty straightforward so far: Four lines (8C/16T, 6C/12T, 4C/8T, 4C/4T) with two or three chips per tier.

What do you mean with P2 btw? To that either means 80286 or Pentium II, neither makes sense in the context.
They weren't all released at once, there was the Pentium 60 and 66, then 75, 90, and 100 with the 120 later. Once we get to the next set of Pentiums we're also going to see a couple on the Pentium Pro side. Though the 200Mhz Pentium Pro ended up coming with 3 different cache options... so the count is odd there...

Wait, what was this thread about, the Ryzen 3? I'm waiting AMD! If the Ryzen 5, 3, and APUs were out, that would be at least 3 more sales AMD!
Posted on Reply
#22
cdawall
where the hell are my stars
ixi said:
You will know it when it's released, nor it's estimated as tdp.
That response makes no sense inline with what I said.
Posted on Reply
#23
medi01
I find the whole TDP thing Huang level confusing.

If chip rated at 65w is actually using up to 150W, another chip, rated 125W, also the same, what am I supposed to conclude from it?

Computer base mentioned high power consumption as cons of 1600x, by the way (they did bother to undervolt it and it did save about 30w)

TheLaughingMan said:
Well the market expanded, needs diversified, and pricing segments got smaller but increased in number. Its the shotgun approach. Release something for everyone and see what sells. Adjust your production according to what ending up being your demand areas.
I think it also has to do with "harvesting".
With single core, if core was faulty, that was it.
Posted on Reply
#24
Frick
Fishfaced Nincompoop
Thermal Design Power dudes. Not power dissipation.
Posted on Reply
#25
TheLaughingMan
medi01 said:
I think it also has to do with "harvesting".
With single core, if core was faulty, that was it.
Yeah. AMD likes to lump that in with "binning" which I find to inaccurate. That started during the Phenom II era which is how we got the infamous Phenom II X3 720. It is now an accepted practice for AMD and Intel to harvest as you say chips that don't live up to spec testing. And while true, I do not believe they are creating segments just to cover some of those chips as their yields have gotten very very good in the last few years.
Posted on Reply
Add your own comment