Read more.Quote:
We show you performance and price.
Printable View
Read more.Quote:
We show you performance and price.
Any word on motherboard pricing to go along with those CPU prices? I assume they'll be reasonable as so much is integrated in the CPU now. The fact that Intel CPUs are expensive is only part of the equation - socket 2011 boards are also not cheap!
I just watched a LinusTechTips video about the launch. XFR seems a bit anaemic to me - it just boosts the core frequency an extra 100Mhz providing you have sufficient cooling...
https://youtu.be/3rUndzpdo1I
It's disappointing that the 4/6 core chips aren't available at launch. They will probably offer much better value for gaming performance.
I'm kind of hoping there will be a nice 8 dimm amd ryzen board (3D design so lots of memory is good :)) because that price difference between the intel and amd based on those stats is enough to double the memory of a 4 dimm board AND get another gpu to boost gpu rendering... hell depending on board prices I can likely build another rig to make rendering even faster via distributed rendering etc.
If there isn't an 8 dimm board then to be honest there will still be a lot of people going intel in the 'professional' field due to ram constraints. I'd like to think intel will now lower the prices of their 'top tier' cpu's at the very least.
Still waiting on a mini-ITX AM4 board before I can price up a system
The Cinebench numbers listed for the Intel CPUs are a little lower than I have seen elsewhere, Anandtech for example list a score of 1547 for the 6900K. Despite that a very encouraging numbers for AMD, just a shame it is only a single benchmark they have released... I really hope they didn't cherry pick something unrepresentative, that would be an awful marketing gaff if it comes to light.
Can't wait to read the proper reviews and see if those performance claims hold up, also can't wait to see the pricing and power characteristics for the whole platform, looks like it could be solidly cheaper than an Intel equivalent using i7-6900K and around the same as an i7-7700K build. Hopefully AMD haven't screwed Ryzen with a power hungry chipset, terrible idle power consumption or some other hidden handicap...
Fingers crossed for competitive choices in the market at last!
Apparently, there will be 82 boards available at launch, so hopefully ITX will be part of that.
I'm mATX so I have already seen a few options :)
Gonna have to get a load of popcorn in today because the whole interweb is going to kick off. Pity there is no real info on the boards available, from what I can pick up from rumours, the MSI X370 XPower Gaming Titanium seems to be the kiddie due to the power delivery system with Std ATX, EPS 8+4 and an auxiliary 6 pin for PCIe. It also has dual M.2 + U.2 support. Normally I am a Asus fan but this board has me thinking.
Another dependency will be who EK Waterblocks designs blocks for.
I am excited about Ryzen except for the fact of, no support for Windows 7. I'm one of those people who are not ready to use Windows 10 full-time. I have a dual booting right now and I'm slowly getting used to it. But I'm not really happy with.
If the legally promised full support till 2020 they should provide that especially to full retail purchasers. OEM users are on their own
The numbers listed by AT are with a custom cooler - AMD is most likely using the Intel stock cooler(it exists for socket 2011 BTW).
I saw this mentioned on another forum:
Also as you know Ryzen is an SOC,so the chipset only really functions as a port multiplier of sorts.Quote:
According to the slides, the i7-6900K scores 162/1474 single/multi threaded in CB r15. Per Anandtech, it does 153/1547.
I think the score discrepancies are due to a combination of two factors:
1) the Anandtech review explicitly tests with Turbo Boost Max 3.0 off, and
2) the cooler used in the review is a closed loop water cooler
With TBM3 off, the max single core clock is 3.7 GHz. If it were enabled, the chip could boost to 4 GHz (which would push the score to 165 if 153 were obtained at 3.7 GHz). And in past tests, AMD has claimed they are testing the Intel chip with "stock cooler." If you've ever seen Intel's 140W rated cooler you will instantly see that it is not up to the task. So when fully loaded, the chip is likely throttling back from its usual all-core turbo.
So I think the numbers are plausible, and it is based on enabling TBM3 and using Intel's terrible reference cooler with the 6900K.
It must have a ceiling, though, because it has to know if the XFR frequency is stable and if the motherboard can cope with the power draw.
If Ryzen cores was capable of more than 4.1-4.2GHz without major increases in heat and/or voltage then surely they would release 4.2GHz quads to fight the i5 7600K. Instead, the quads cap at 3.9GHz standard boost (according to the so far very accurate leaked list) and only the most expensive cherypicked octocore is offered at 4GHz standard boost. I'd actually say this is the biggest problem with the lineup; there really needs to be a 4GHz hexcore to fight the 7600K and 7700K near their price point but the hexcores are presumably low frequency to scavenge as many defective octocores as possible.
Lower frequencies are probably the cost of having such a compact core (and not using Intel's better fab process). I still think this gives good enough single threaded performance to justify the extra cores and/or lower cost they can offer.
fantastic. we finally have a reason for intel to get up off their fat arse and try and do something. Hope we see lots and lots of people buying these CPUs, they look great.
Now we see why Intel's monopoly of the last years has been so bad for consumers...
cant wait :) maybe this will lower intels cpus and MB !
Yep I'm waiting for Ryzen to reduce Intel chips for me. Had AMD in the past and they just don't quite deliver as well as Intel.
So basically you are emotionally tied to Intel as a brand instead of looking at chips on their own merits??
What is funny is AMD CPUs have been used in supercomputers by large aerospace firms and in academia in the UK and yet people still expect AMD to drop Intel prices for them??
So I expect you were one of the people who bought P4 CPUs on the desktop when AMD had the Athlon XP and Athlon 64 out??
Its sad when I started as an enthusiast,people bought what was best value for money,and had no problems buying AMD CPUs like the Athlon XP and Athlon 64.
So AMD probably has something decent and you continue to buy Intel - no wonder gamers are paying through the nose for things. If AMD has something decent and they can't make money,then they might as well not bother making CPUs for us at all.
No wonder Intel hardly drops prices when they have a captive market who will only buy their products.
Reap what you sow.
Edit!!
If you use your logic its like saying somebody would never buy an Intel CPU since they had a Pentium 4 in the past.
I'm thinking that Ryzen looks extremely attractive !
It is but sadly PC gamers and enthusiasts are getting more and more tribal like with console gamers. This is the problem - AMD needs to sell enough CPUs,and if Ryzen does not do well they are screwed. The problem is many are just using all that R and D money AMD used for Ryzen to fund their purchase of Intel CPUs,not trying to look at the products themselves on their own merits.
Many seem to make lame excuses forgetting Intel has had rubbish CPUs like the P4,Itanium and Atom too.
It was the same tribal mentality where you had people who would never consider a Phenom II X4 over a more expensive Core2 quad,just due to the name.
Intel is the same company which recently released the £175 Core i3 7350K yet go back 5 or 6 years ago and we had the sub £100 Core i3 530 which was could be overclocked.
Its like stealing from the poor to fund the rich.
I really hope that AMD does better in OEM and commerical sales - the gamers and enthusiasts of today are not the same as 15 to 20 years ago.
Its less bang for buck,more like brand for buck with DIY builders nowadays.
ryzen finally puts amd back on the map i'm actually considering buying myself one, might wait to see the price of the MB first tho
can't wait might buy one i'm going to wait for the benchmarks to come out but its looking good so far
Ebuyer has their preorder page up: http://www.ebuyer.com/search?qfc=778677|778679|778680
Priced at £320, £390, £490 so pretty much exactly in line the the US$ prices + VAT. 8C/16T chips for £320 ... what a time to be alive ;)
Oh, and also an AMD processor that they think they can sell for almost £500 ... it's been a *long* time since you could say that :o
Its looking like the 1700 is the one for me, just waiting on reviews and overclock results.
Amazon have the same pricing with a pre-order guarantee that if the price goes down you will only pay the lower price.
I am going to keep an eye on Ryzen, as I used a lot of 3D apps and currently have an i7-4790k it could well be my next processor.
However will wait for real world benchmarks etc and see what the state of play is then
Just spec'd an upgrade on scan. £1200 :( i was hoping the chips and boards would have been slightly cheaper. Damn the exchange rate. Still, 8/16 with good single core for that price. Kudos amd. Lets see what the benchies reveal in the coming weeks though before i kill my wallet. Ill need max oc figures.
Scan have prices for quite a lot of AM4 motherboards:
https://www.scan.co.uk/shop/computer...2496/2497/2436
Well Ryzen made it...took forever
BitWit
https://www.youtube.com/watch?v=qTu7ryfKO5E
Paul's Hardware
https://www.youtube.com/watch?v=QlWbpUDOigU
Official video of the 24 minute launch event:
https://www.youtube.com/watch?v=1v44wWAOHn8
Very interesting, don't really need an upgrade but I do want to get an NVMe drive and I don't think Z97 as a platform makes the most of it due to not having PCIe 3.0 x4 otherwise I would have been more than fine.
How come I am seeing HDMI, VGA and DVI outputs on motherboards? Presumably for future APUs?
For existing APUs, more to the point. Bristol Ridge APUs for AM4 are already available to OEMs ... whether we'll see them in the channel or whether AMD will just hold out for .. erm, Raven Ridge, I think (getting lost with the code names) ... anyway, for the Zen APUs later in the year.
Personally, I hope we get a range of Bristol Ridge APUs released alongside the motherboards, so there's something to fill out the £100 and lower range with the latest technologies. Given Ryzen goes down to full quad cores which will be a LOT faster than even the last generation Excavator cores in Bristol Ridge, it'd make sense for desktop BR to fill the lower end of the range.
Not just future, the idea was that AM4 is the unified platform and even Bristol Ridge APUs (Excavator) uses AM4:
https://en.wikipedia.org/wiki/List_o...16.2C_28_nm.29
However, rather than having a flood of AM4 boards out way before launch, it looks like Bristol Ridge only had limited traction*.
Of course, for now, the video ports are not going to get used until Zen APUs come later in the year.
EDIT: *Or as scaryjim said, it was OEM only.
If you want to get an idea of the obtainable overclocks (and can't wait for the reviews), scan are selling factory overclocked PCs with the 1700 @ 3.8GHz and 1700x @ 4GHz and oddly 1800 @ 4GHz
Well, the initial release certainly was. I'm pretty sure AMD claimed they'd be released in the channel eventually, but I have a feeling that most mobo manufacturers were unhappy releasing AM4 motherboards when the only processors for them would've been significantly weaker than anything Intel was offering. AM4 boards make a much more enticing case when there's a set of very competitive processors to go in them. Now the boards are released, AMD might release Bristol Ridge to the channel so there's a full range of processors available. Or they might sit tight until Zen APUs are available for the channel: otherwise there'll be a lot of FM2+ products in warehouses that are never going to sell...
It seems you can overclock individual cores:
http://i.imgur.com/ZwY05FK.jpg
http://i.imgur.com/ZwY05FK.jpg
You can say the same about any new standard or technology.
The 5 gigabit standard was approved precisely so that there was an easier and cheaper transition to faster ethernet for home networks. I also recall reading somewhere that 5 gigabit switches and maybe routers are in the pipeline for this year.
It seems a shame to have a new start with Ryzen yet not marry that with the new standard for LAN (except, as far as I can see, the ASRock Professional Gaming - which also looks hideous, but I'll stick it inside a Corsair 600q if I take the plunge). After all, 10G has been sporadically available for Intel's HEDT for a while now.
Well, looks like it's going to be an AMD system, my next one; although ideally I'd like a dual Ryzen system.
Its only feels real when you can put then in your PC parts Picker Builds...so now you can...the universe feels safe.
I had to actually both (2.5G and 5G) of them up as all the ‘normal’ Ethernet standards have been base 10
Both came long long after 10GBASE-T and it seems for one reason only: 2.5GBASE-T can run on old existing Cat5e cables so the idea is that a business could upgrade some of their network to that without the expense of upgrading all the old cables to Cat6 or Cat6a.
However, 5GBASE-T really makes no sense whatsoever as it needs Cat6 cable (not ‘a’ so good for 55m at 10GBASE-T) so why not go all the way to 10G in the first place?
Unless 5G hubs, switches and routers are going to be a lot cheaper than 10G it really makes no sense. 10G stuff has been around for ages so has a had to come down whereas 5G is all new.
Either way, nothing for home users where existing cables are seldom a concern.
Yep, that is the expectation; 10G switches after all these years are still very expensive, whereas 5G switches are expected to be much cheaper and should fit in a home premium sort of price bracket.
The cost of Cat 6 cables for 5G ethernet isn't really a significant factor for home users since few homes have long runs of cabling (most would use powerline ethernet or wireless instead - and mostly due to the expression on the wife's face as you start drilling holes in the wall rather than the cost). Hence, as you pointed out, 2.5G over Cat 5e is more for businesses.
When will they be releasing Ryzen APUs?
2.5G is a nice multiple of 25G, which in turn is a nice multiple of your 100Gbit backbone network ;)
40Gbit seems to be getting popular, but is still expensive so I gather people get a pair of 25Gbit ports for the same price giving them 20% more bandwidth for free.
But for most home users, they are on wifi with miserable throughput and reliability. Gigabit is way more than they have and more than they realistically need. Perhaps when network chips start getting made on 10nm silicon it will become cost effective to have faster ports as it won't cost much more to make, but I'm struggling to see a demand from the average user.
Anyone got a comparison between the AM4 chipset's, I'm struggling to find one or just haven't had enough coffee yet
HUh, I'd swear Hexus had those slides in an article somewhere but I also can't find it.
Here's the relevant slides on PC Perspective: https://www.pcper.com/reviews/Mother...-AM4-Continued
The chipset comparisons for AM4 are somewhat muddied because the chip that sits in the socket is actually an SoC, rather than just a CPU - as well as the memory controller and PCIe lanes (16 Gen 3 for Ryzen) it includes a southbridge that supports a variety of SATA, PCIe and USB configurations. So the exact combination of IO available depends on both the chipset used and the CPU/APU in the socket. AMD have even confirmed that motherboards without a chipset - described as X300 or A300 (depending on whether overclocking is possible) - are allowed, and these will derive all their IO from the CPU/APU (which will make for some very clean, simple and hopefully cheap mITX boards!).
Cheers Jim appreciate that. It was the SATA I was interested in as have 7 sata devices but currently see a lot of boards supporting 6 and looks like its just the top end chipset that supports above that. Well the ones I have looked at, I do have a SATA add in card but trying to reduce the amount of cards in the pc currently.
A lot of boards manufacturer appear to have gone for both an M.2 NVME PCIe slot and an M.2 SATA slot. If you can find one with 6 SATA + SATA M.2 (The ARock X370 Killer SLI seems to offer this), you can get cheap passive M.2 -> 7 pin SATA adapters that would allow you to use 7 standard SATA devices...
B350 and A320 motherboards could technically support up to 8 SATA devices if both SATAe ports were dedicated to 2 SATA 3 ports and the CPU/APUs storage controller also provided 2 SATA 3 ports. In fact, they could also support a PCIe NVME at x2 in that scenario, so the possibility is there (should any manufacturer feel so inclined) to offer a storage focused board that could take an NVME SSD as a boot drive and 8 SATA drives (although I'm not entirely sure if you could raid across the whole array or if the two drives hanging off the [CA]PU would be separate).
EDIT: The ASRock AB350M Pro4 also offers 6 SATA3 + M.2 SATA at a more palatable £94...
Thanks Jim, going to look into that then I think. If the benchmarks pan out and I can see a real improvement over my i7 then will be moving to the AMD platform.
If money's no object, ASUS' just announced Crosshair VI HERO has 8 SATA 3 ports natively (and no display outputs, interestingly). Mind you, it will set you back £260... :eek:
EDIT: turns out the Prime X370 Pro does too - that's a far more reasonable £160.... :O_o1:
Yeeeeaaahhhh think I will probably be avoiding that one :)
Edit : And the other one lol
You can add sata cards from as little as £5.....no point spending through the nose for a couple of extra SATA connectors....
Thats what I have now, a PCIe Sata card but its right up next to my 970 as the other slot is occupied by an Intel nic so was trying to see if there was any better options with the AMD but seems 6 sata on low/mid range boards seems the norm.
However from looking at it I may get away with not needing the Intel nic anymore so could move the sata card to their I guess
http://i.imgur.com/40l2GhV.jpg
It seems the Ryzen 5 1600X is a 3.6GHZ to 4.0GHZ part and not a 3.3GHZ to 3.7GHZ one as leaked before!!
That sounds more reasonable, actually. Presumably the leaked spec is a non-X part, or maybe a 1500X? (like the R7 has 1800X, 1700X and 1700). That'd essentially make the 6 and 8 core variants 95W processors with low-power variants (1700 and 1500), while the 4 core parts are inherently 65W.
https://www.dvhardware.net/news/2017..._leak_wccf.jpg
That is the leak - the top three SKUs were accurate.
Yeah - the if the one listed as as the 1600X is actually a 1500X, the 1600X from your slide just drops in on top of them (at $289, maybe?).
That'd make the 8 and 6 core range 1800X (95W), 1700X (95W), 1700 (65W), 1600X (95W), 1500X (95W), 1500 (65W). That'd make a lot of sense, I think...
I am hoping AMD gets to at least Haswell level IPC in gaming on average,because that is going to be a very competitive range there.
So unless the pound gets weaker,a £200ish 4C/8T Ryzen with 4GHZ XFR boost clockspeeds in a mini-ITX motherboard will be a nice upgrade over my Xeon E3 1230 V2 and I know I have an upgrade path to an 8C/16T CPU and probably a greater chance of having long-term motherboard support for my CPU too.
Its kind of why I want to move away from Intel - my original motherboard was a higher end H67 one and when it went kaput,Intel had moved to Haswell and I had to get a lower spec B75 based one as the only two mini-ITX motherboards with a reasonable spec available in the UK by that time were B75 based ones.
I think anyone with a current AMD system will see an enormous jump. Indications are IPC is equal to Skylake in some benchmarks. In fact, in the video posted earlier they stated that they exceeded the 40% IPC goal and got 52%. Can't wait to see how prices settle and Intel's reaction.
The CPC leaked of the qualification tested six games with a Fury X. Four of them only scaled to four cores at most and did not really use SMT that well,and the other two tend to need at least 4 decent cores and could use up to 8 threads. The score hinted at close to BW-E level IPC. I am only estimating Haswell level to be more conservative.
Just priced an upgrade.
ASRock X370 K4
1700
32GB ram (2x16gb to leave room for later 64gb upgrade if needed).
Comes in around £700 and will low my current system out of the water.
May be worth doing sooner rather than later while my 3570k has some value before the R5s make it worthless.
iirc my last calculation said 92% of Broadwell E IPC for the main cores, but 97% when SMT kicks in (AMD's SMT implementation appears to be very efficient, at least in the workloads we've got leaked benchmarks for).
That means on current leaked clocks AMD are going to have a narrow overall lead in HEDT, where Intel's chips have lower clockspeeds (and much higher TDPs, incidentally).
In the consumer space it's a bit less clear cut because Intel's chips clock higher (and of course the popular-with-enthusiast K chips have higher clocks than the non-K chips). For instance, AMD's 4C/8T 3.6GHz - 3.9GHz+ R5 1400X will probably get compared to the i7 7700k, which is clocked at 4.2GHz - 4.5GHz. That's going to give it a > 10% clock speed advantage, as well as a ~ 10% IPC advantage, so it's going to come out handsomely on top. However, that's a £330 chip compared to a £200 AMD chip. The ordinary 7700 clocks at 3.6GHz - 4.2GHz, so it'll be a LOT closer between that and the AMD chip, and if you go to Intel's £200 you're looking at an i5 7500, which clocks at 3.4GHz - 3.8GHz, a couple of hundred Mhz behind the AMD chip (and without the option to overclock, incidentally). That probably won't be enough to put AMD ahead in single threaded tasks, but it will make it very close, and obviously having SMT the AMD chip will easily win any multi-threaded tests.
Ultimately though, the end result is likely to be that differences between AMD and Intel CPUs are negligible in all but a small number of outlying edge cases.
The real question for AMD - as ever - is not winning over the enthusiast but getting Ryzen adopted by the big OEMs. The performance looks like it's going to be thereabouts, and the pricing is definitely right, but without significant OEM market penetration it's going to be difficult for AMD to shift massive volumes of processors - after all, the majority of PCs are bought pre-built from big name integrators, not piecemeal for self assembly.
I just searched 'Ryzen' in Amazon, top result is an Intel i7 7700k, followed by An AMD8350. The proper Ryzen chips don't appear until about half way down the page. And this is sorted by relevance, apparently. Think someone has been paying Amazon to skew their search results...
Well, a lot Intel's crazy market segmentation will now come back to bite them.
Most likely Intel are going win at AVX2 loads too but luckily for AMD AVX2 software support has been very slow since Intel nicely kept using AVX to segment the market (because fusing it off is a lot simpler than having considered how to introduce wait states during design).
So that's Intel's market segmentation might work against them. Which is nice.
The new artificial segmented chipset for each CPU has a major inconvenience for the DIY market for years now.
Max frequencies do seem a problem for now, but what we've seen so far would be ideal for servers and big server buyers do actually pay very close attention to benchmarks. If Naples takes off, I wonder if any of the very big data centres would be willing to order a semi-custom part with some specialised fixed-function hardware.
That new hexcore slide Cat found goes on about Q2 so maybe those will be from a re-spin? Don't know what the lead time for the process is, but if some the initial ES's and what they had back in December had an issue the managed to fix for max frequency and they had to re-spin it Q2 for hex and quad makes sense.
Maybe the initial stepping was okay running the octocore at the speeds we see now (which are actually quite good for a 8C/16T chip), but for quad they need more.
Nah, Amazon have one the worst search engines where anything which has a comment with the word gets a hit. Almost no filters nor any intelligent search keywords (AFAIK). At one point ebay took away their OR (which you get by using brackets and comma separate list like (1, 2, 3) and so on), but they put it back. Plus they have tons of filters too.
Crazy to think such a poor website like Amazon is one of the world's biggest retailer.
That's the ticket!
IMO that's the perfect balance between gaming, productivity and price.
I'm loving that the totally artificial segregation of consumer and professional Intel has been driving all these years with Z*70 and X*9 is being flattened with AM4. The constant socket changes have also been a real drag.
If previous AMD chipsets are anything to go by, AM4 should stick around for a while.
Think my next upgrade will be a Ryzen system
;)
I'm going to wait until Vega has settled in, then maybe get a complete AMD system. The highest level Ryzen chip I'll consider is the Ryzen 7 1700. You never know, there maybe bundle deals with a Ryzen chip, a Vega card, a mobo and some ram etc even if it's just a retailer doing it, although I'm thinking AMD will be considering such an offer too.
Timely article over at ComputerBase:
"6, 8 or 10 core CPUs beat 4 faster cores"
https://www.computerbase.de/2017-02/...e-spiele-test/
Nicely done in this review is that you can select only the games you want (click on 'Bearbeiten' and uncheck the ones you don't want).
https://i.imgur.com/li4fSen.png
Most of the games are fairly recent which helps. With the amount of people doing video recording while playing or doing other things in the background the real world benefit of more cores even for games is quite high.
https://s4.postimg.org/4hzyydz0d/Scr...223_172714.png
DinoPC just leaked that - its comparing a 65W TDP R7 1700 against a 91W TDP Core i7 7700K.
GTA V tends to not scale well beyond 4 cores or use SMT well:
http://www.techspot.com/articles-inf...nch/CPU_01.png
http://media.gamersnexus.net/images/...u-1080-max.jpg
Interestingly enough the difference in minimums seems similar to the difference in base clockspeeds of each CPU,ie, 3.0GHZ to 3.7GHZ for the AMD CPU and 4.0GHZ to 4.5GHZ with the Intel CPU.
I suspect the AMD CPU is throttling due to TDP,but if that is the case the 95W TDP R7 1800X and 95W TDP R5 and R3 CPUs,might actually do better than we think.
Also,a bit of info from Gibbo over on OcUK about the leak:
I suspect that hints that even the R7 1800X might be somewhat limited by its TDP.Quote:
I'd ignore test, 1700, base is 3.0GHz, turbo is 3.7GHz, AMD do this so it qualify as 65W part. If you switch it over to manual mode in a good motherboard like Crosshair you just lock it to run at 4GHz all the time, then the performance is vastly better, even beats 1800X. :)
Edit!!
Apparently in CB R15 the chip scored 153 points at 3.4GHZ,which places it close to BW level ST IPC,but SMT scaling seemed better.
This matches my thoughts, rather than get the chip with the extra clocks and boosts get the cheapest and overclock, like wise if it can boost to 3.7ghz there is no reason it can't stay there and probably go higher with overclocking.
I hope these chips will be tested on win7 since 50% of us use it. These chips have no gpu and AMD only mentioned Bristol Ridge having issues (as they won't be writing gpu drivers SUPPOSEDLY for these). AMD won't be selling many chips to DIY people if they try to force win10 on us...LOL.
We know you can run Kaby Lake on Win7 per anandtech's review:
http://www.anandtech.com/show/10968/the-intel-core-i7-7700k-91w-review-the-new-stock-performance-champion/2
Kaby on Win7 SP1 at anandtech. No use of the gpu (or mention of it) but you can get drivers here if needed for that chip:
http://www.station-drivers.com/index.php?option=com_remository&Itemid=353&func=startdown&id=2469&lang=en
Intel HD & Iris Graphics Drivers Version 21.20.16.4526 WHQL, a forum poster tested it. Can't remember which forum, I just kept the files and links...LOL. Just in case I go Intel instead of Ryzen. Either anandtech didn't know about the drivers or WINTEL told them to ignore them and act like the gpu won't run in win7. Who knows.
For anyone interested:
The v21.20.16.4528 also say the same, but haven't seen someone test them.
http://www.station-drivers.com/index.php?option=com_remository&Itemid=353&func=fileinfo&id=2514&lang=en
All of the drivers after this (09/23/2016, posted oct) say win10 for gpu on kaby. So I guess they supported it up to a point (probably for enterprise or something).
It will be interesting to see if AMD makes a win7 driver to sell more chips. They could sell a lot to enterprise if they made drivers for integrated versions coming later this year. Tough to tell a large business to move all their users to win10 when they just went to win7 and planned on staying there for another 3yrs...ROFL. Good luck Intel/MSFT/AMD. Forcing business to go where they don't want to didn't work out well with XP, and I suspect it won't work any better with Win7.
I expect they will work on Win7, they will probably not be optimised on Win7 though so things like power saving might work in some old-school compatibility manner. In which case, that would be good enough for most people. I'm sure we will find out pretty soon.
What with talk of all Ryzen coming with XFR I'd take a wait and see approach as if it's true the X chips must have something else different about them, my guess is that although all Ryzen chips are multiplier unlocked and come with XFR that only the X chips allow per core adjustment of settings along with disabling individual cores.
If that turns out to be the case i expect it's only going to work on the fly with Windows 10, it would also give Ryzen a selling point that couldn't be matched by a simple management decision from Intel to drop prices.
Erm ... where've you heard that? AFAIK XFR is only enabled on the X chips - that's the difference between them and the non-X chips.
All Ryzen are unlocked for overclocking - so you can manually bump the speed of any Ryzen CPU to your heart's content. But that's nothing to do with XFR, which automatically boosts the clock speeds as high as your cooling can cope with. Essentially it removes the processor's configured TDP limits when calculating how high it can boost.
Also AFAICT, XFR is entirely driven in hardware, so it should be OS agnostic - it's not dependent on having an appropriate software driver or anything. I think making it a Win 10 only feature would be fairly suicidal from AMD, given that they're pushing the chips quite heavily for creative professionals as well as home users/gamers/enthusiasts...
I heard it from hear, hear, and hear, maybe they've got it wrong and there's just confusion but it seems odd that two separate retailers and potential reviewer would say the same thing.
Yea XFR is hardware driven but AMD's Ryzen master software has show we'll be able to tweak and disable individual cores, that's what I'm guessing will be Windows 10 only.
at last I can start planning to upgrade from my Phenom II Hex core
a 1700 with 65W is a beauty I think
Whether they've got it wrong or not, there's definitely confusion (if you read on through those thread, particularly the bit-tech one). Perhaps XFR is available on all processors but disabled by default on non-X processors?
That'd make more sense to me than the X indicating per-core adjustments that aren't available on other chips.
This whole thing about XFR being limited to very small boosts also makes no sense whatsoever when you look at AMD tech slides about XFR, which explicitly say that frequencies scale with cooling. If the max XFR boost is 100MHz that'd patently be untrue.
The software looks to be based on the same interface as the existing Radeon settings software (which would make sense), which AFAIK is supported under Win 7 - 10.
OTOH, AMD announced a couple of weeks ago that they'd only be doing Ryzen supporting drivers for Windows 10. Took a bit of hunting, because I knew I'd seen a story saying they were supporting Windows 7, but apparently there was a clarification a few days later. AMD have validated Ryzen on Windows 7, but won't be supporting it.
http://www.pcworld.com/article/31672...windows-7.html
So yeah, software malarkey for tweaking overclocks will only be available on Windows 10, I guess. Wonder if any mobo manufacturers will be implementing per-core adjustment in UEFI....?
It's a bit like trying to guess what's inside the box sitting under the Christmas tree ain't it. :)
Do we know where X=XFR came from, was it just an assumption?
Reading around it must be in an AMD marketing pack that's gone to retailers somewhere, since most retailers and AMD's own pre-order page list XFR as a feature for the X versions and not for the non-X. The other information appears to have come from the engineering side and deeper conversations with AMD directly (not sure where Gibbo got his info, mind you!), which would imply to me that all the chips are capable of XFR (unsurprising, since they're the same silicon ;) ).
Until I hear otherwise, I'm sticking with my assumption that the 1700 has XFR disabled but it's possible to turn it on, whilst the 1700X and 1800X have it enabled as default. ;)
You think it's possible that this will be an unlocking-shaders-esque situation where later revisions make enabling it impossible? What's the benefit of an X version if XFR is available to non-X? Can't imagine "we made sure these chips were lottery winners" is a good enough reason for £80 more, nor is it something you can officially advertise IMO.
Does this mean Moore's law can be reignited. Although Intel had problems getting beyond 14nm for quite some time they've also been given plenty of time to work things out. This looks set to be a very competitive 5 years.
Again, that is an optimisation. Piledriver worked just fine under Windows 7 before the scheduler patch came out, which got you another 5 to 10% performance not non working vs working.
Now I would hope that the existing scheduler support for hyperthreading would work just fine on Ryzen and there isn't much to be had. It could possibly be taught when best to migrate threads vs when to sleep cores and use the power elsewhere, there might be mileage to be had from knowing that the chip is made as two banks of four cores so there might be L3 cache hitrate improvements from dividing work up in a NUMA fashion. First is a power management issue, second is the level of finesse that will only be delivered in the latest version of the OS where all the main work is happening. That is the difference between "validated" where it will run and "supported" where it will run as well as possible.
That isn't just Microsoft, I wouldn't expect Red Hat Enterprise 5 to get full Ryzen support despite still being an officially supported OS, it just gets bug fixes these days and that is a feature. :)
I don't think Piledriver was a Moore's Law blip any more than the Pentium 4 was caused by lack of transistors or transistor performance. I still believe AMD could have built a 6 core FM2+ chip with some L3 cache and it would have been way more competitive than Piledriver on AM3+ but clearly their heart just wasn't in the construction series CPUs any more.
From OcUK forums:
I heard about that website - a few years ago they actually used to be quite good at leaking stuff before the likes of Wccftech.Quote:
From one of the biggest Hardware publications in Turkey apparently.
Saw this video earlier but I couldn't understand what was being said. However someone has translated some of what was said thankfully.
This is the video; https://www.youtube.com/watch?v=UhKmeCdB914
To sum up the things he stated;
"Overclock is no problem at all. Though we can not say a certain frequency due to NDA, I can say that it is impressive. Overclockers will be happy."
"I have friends from AMD Engineering department due to my experience in the hardware industry. Zen 2 or Zen B will be even more competitive. So we can say that Intel shall brood on the future"
"With our overclocked 1800x sample,under Noctua cooler given by AMD, we have passed beyond the stock single thread performance of 7700k, in a specific bench, and the temps were great. We had no concern about temps during our run which passes the ST performance of 7700k."
"In some benchmarks AMD(!). It seems ironical yes but AMD is presenting a CPU performance that Intel can not keep up with even with their 10 core 6950x!"
"Breaking NDA won't be a problem since the scores are beyond fantastic. But we will stick with the tradition."
"With one click I can reach great OC's. So I won't really bother with the manual OC no more."
"Single thread score will be so great. According to this performance we can say that 7700K will be history, even for gaming, from now on" he said.
"Intel shall shake themself. Because Ryzen will be a great choice for Overclock enthusiast."
Edit!!
From the reddit thread on the YT video:
https://www.reddit.com/r/Amd/comment...ation_to_fuel/
Quote:
That's roughly what they say in the video. But I would like to clear up some things.
I don't think he says he won't bother with manual oc anymore. He says "I can reach a fairly nice overclock with one click. You don't need to bother with manual OC"
He explicitly states that SC performance is still behind 7700K, but with OC and in one specific test it passes even 7700K in single core performance.
He also says with OC, the CPU has simply no competition and that not even 10 core 6950x can keep up with the Ryzen when OC'ed.
By the time Star Citizen actually releases we will be on 16 core CPUs ;) :D
Lol :) I have had Star Citizen for a while now, they are releasing a lot of things but personally I think they are spreading themselves a bit thin, would rather have seen the main game come out then the extra's done afterwards. Time will tell.
I know when it does come out I will be on the Ryzen III - 48 core
Looks like the Core i7 6900K is dropping to $699:
https://twitter.com/BitsAndChipsEng/...85680554196992
Has amd finally made a cpu that's superior to Intel?
Sent from my Nexus 6P using Tapatalk