That's suspiciously close to the 10% that was ballparked for zen+ previously, but for IPC alone. I wonder how much is down to faster RAM? I'd expect the 2X00 CPUs to at least match the 2.9 GHz rated RAM speed of the APUs
Printable View
Interesting .. given the relative clocks that's indicative of a reasonable IPC uptick as well (around 8%, if my maths aren't way out). I'd assume the big jump in multithreaded score (+30% :O_o1:) is down to Precision Boost 2 being able to run the cores much faster in the MT benchmark.
I suppose the big question is whether the 95W SKUs are going to see similar improvements or bigger ones. If the 12nm process has a flatter v/fmax curve, *and* it can do +200MHz in the 14nm sweet spot, we could see something pretty stunning....
Looks like AMD has hired a senior person from Cisco:
https://www.theregister.co.uk/2018/0..._epyc_new_gig/
Edit!!
CPU tests on RR with faster RAM:
https://www.phoronix.com/scan.php?pa...Memory-Scaling
It seems to have better CPU performance with faster RAM,so it appears even a single CCX design likes faster RAM(unless OFC it is down to the lower amount of L3 cache).
One of the major arguments for Ryzen needing faster RAM was down to the bandwidth between a pair of CCX modules being limited,and faster RAM would mean less of a cross CCX penalty,when using a second CCX for CPU stuff.
OFC,I would love to see some gaming results with a graphics card,to see if faster RAM does impact CPU performance significantly,but apparently most sites have not bothered doing this. Intel on average gained less from faster RAM.
Part of the assumption about memory clocks boosting Summit Ridge performance was that it sped up communication between the two CCXes, because the fabric clock is tied to the memory clock. On the surface it makes sense: you speed up the access to the other CCX's L3 cache, which increases performance.
Thing is, most people ignored the fact that cross-CCX cache latency was pretty much identical to main memory latency - SR CPUs essentially acted like 2 processors with 8MB L3 cache each and a shared main memory access. Speeding up the memory does reduce the latency of cross-CCX communication, but only to the same extent that it reduces latency back to main memory too (i.e. there's no gain from having the data in the other CCX's L3 cache). So it's not really surprising that you get similar gains from faster memory on RR - you're reducing the latency associated with misses in the L3 cache.
One notable thing about the Ryzen IMC is its very high latency to main memory - having a quick look at the MSI Mortar AM4 board review, Coffee Lake with 2400MHz CL 16 RAM has significantly lower latency than Raven Ridge with 2933MHz CL 14 RAM (as in 13ns - almost 20% - lower)*. That's pretty ridiculous - faster RAM with tighter timings but much lower latencies? That bottleneck has to be in either the IMC or the fabric....
*you get some very interesting results if you look at the theoretical latencies in that review. 14 cycles at 2933MHz gives a theoretical 47ns, v 70ns measured latency. 16 cycles at 2400MHz gives a theoretical 67ns, against a measured latency of 57ns. Somewhere along the line something's gone fishy with the Intel latency testing.... :O_o1:
It will be interesting to see if Ryzen+ improves in this regard. Currently being stuck with 16gb of 2400mhz DDR4 does kind of not help my own circumstances when it comes to my next upgrade.
Sadly you can't keep the boot kits. :( ;)
https://cdn.discordapp.com/attachmen...222_135458.png
PS. when will spoiler handles be a thing
It will be a retail A6 9500:
https://www.reddit.com/r/Amd/comment..._kit_from_amd/
Yeah, I know. They told me all about it in the rest of the email(s). I would have appreciated being able to hold onto it though, which you can't, as they went them back. I can't imagine that AMD doesn't have enough 9500s to ship out to us weirdo early adopter APU builder (with no other Ryzen CPU) types though, so why they want them back is a little beyond me. Can't be worth that much.
Well they are second hand now, can't be sold as new. Makes sense to cut down the number that get used for this, though in the end if they boost the AMD market share and sell some more motherboards that probably isn't a bad thing. Perhaps they will all just end up on ebay :D
Presumably prefetching interfering with their testing methodology? That's just reminded me of a rant I never had. :P Testing latencies is a strange one - I remember reading some people complaining about some software being 'optimised' for Ryzen after initially poor results. They didn't seem to get that there's literally one correct answer for memory/cache access latencies and measuring it correctly may require some software adjustments to allow for some architectural differences e.g. prefetching algorithms. I wonder if said people would also complain if the Intel results increase if they get fixed. Maybe they'll be too correct then?
I wouldn't be overly concerned about it personally, I also have 2400 DDR4 and it seems like, while memory benchmarks and some synthetics show major differences, they're generally fairly minor for actual applications. That's for CPU performance of course - GPUs obviously love as much bandwidth as you can throw at them!
It depends on the game,and guess what game scales very well with RAM speed. You guessed it!!
https://cdn.alza.sk/Foto/ImgGalery/I...ut-4-1080p.Png
https://static.techspot.com/articles...ch/Fallout.png
Having said that it makes me wonder whether the high memory lateny with Ryzen might be the issue here too.
Seems Lenovo are going to release two ThinkPads with Raven Ridge.
The E485 and E585 which are based on the 14" E480 and the 15.6" E580 respectively.
So that means at last a Raven Ridge laptop with a TrackPoint
https://i.imgur.com/EnwGEfK.png
(Source: https://cloud.kapostcontent.net/pub/...IfjbsvUQDKs1jA, which seems to be where NBC got their info)
https://www.notebookcheck.net/Lenovo....285214.0.html
https://www.computerbase.de/2018-02/...e485-e585-amd/
The ThinkWiki doesn't really cover the E series that well, and only has entries for the Haswell models (E440 and E540) on the English site, and Skylake (E460 and E560) on the German sites:
http://www.thinkwiki.org/wiki/Category:E440
http://www.thinkwiki.org/wiki/Category:E540
http://thinkwiki.de/E460/E465
http://thinkwiki.de/E560/E565
It would seem Samsung weren't being overly optimistic with their single threaded performance claims for the Exynos M3 cores! https://www.anandtech.com/show/12460...-the-galaxy-s9
https://images.anandtech.com/graphs/...2460/95525.png
That's an impressive generational jump, I look forward to seeing the rest of the benchmarks including power consumption!
In other news I see some places (wccf included) complaining about 4/6GB RAM as though it's going to impact performance somehow. Seriously??? And with the current RAM situation, the last thing it needs is for one of the best-selling phones to put yet more unnecessary demand on the market!
I didn't bother looking at the rest to be honest - I was only interested in Exynos performance vs the Snapdragon. :P
It would be really nice if this class of SoC along with a decent screen e.g. good quality 1920x1080 IPS LCD along with decent build quality etc made it to more mid-range phones. I'm really not all that bothered about the 'fancy' cameras, pixel density race, thinness, etc.
The biggest improvement is what you showed - the CPU performance is what tech sites should be concentrating on more. At some point,the phones should be able to replace even laptops for more normal usage,ie,using it as a hub which you can run a larger screen,etc off as some companies have tried to do in the past. The other aspect I am interested in is dropping the power of other parts such as the wireless modules and screen,so we can have more and more phones move towards true 24 hour battery life even if you using things like GPS,internet,etc.
Instead they are concentrating on one of the cheaper parts(which cameras tend to be),and the fact it is doing things which even cheapo 35MM compacts did in the 1990s!! :p It was like with the tech sites fawning over "glass lens elements" in a LG phone last year.Like,why is it such a big deal in a £600+ device - I actually thought they already had glass elements and variable apertures. I just hate when simple features are just marketed as being "revolutionary". If that is the case,I am going to stick with a cheapo sub £200 phone and get a £300 compact which by mobile phone standards will be space age for the next few years!! :p
It looks like this B350 board works fine with ECC RAM:
https://www.reddit.com/r/Amd/comment...rime_b350plus/
Instead of all the marketing gimmicks, I'd sooner have a reasonably-sized phone with a decent battery rather than concentrating on paper-thinness, protroding bezels made of some sort of shock-absorbing material to protect the glass screen rather than something perceived as high quality but vulnerable to damage i.e. glass/metal, decent attention paid to the radios, etc. Some of the Galaxy 'active' phones seem to tick a lot of those boxes (at least they did last time I checked which was a while ago TBF) but they are even more expensive than the base model and seldom available in the UK which is a shame.
Yeah I was aware of the Motorola ones when I was looking but I can't remember what it was that put me off. TBH it was around the time of the Snapdragon 810 so that might have had something to do with it. I dodged that by choosing a Huawei phone which used their HiSilicon 955 chip which fitted with my requirements quite well - decent power management, a focus on useful CPU performance rather than chasing benchmarks with a huge GPU, etc. That, and I had to return two Moto X (2014 IIRC) because of a fairly severe screen issue they both had (dimmer and distorted colours towards one end of the OLED panel and evident screen burn out of the box!) At that point I assumed it was a bad batch rather than a one-off so gave up on it.
Something I've picked up on is that at least some phones do seem to change specs somewhat after the initial batches. A friend had a Galaxy S5 which had considerably faster storage than the original review models!
Are there any phones powered by AMD CPUs? (the original topic :) )
Mate has a Moto X Force and that is one tough phone,but Lenovo software support is meh,and they lied about updates too. Regarding the HiSilicon SOC - mine has the 650 and a mate has one with the 655 and I can confirm both phones have excellent battery life and general performance,and that includes stuff which tends to use GPS a lot.
When have 'AMD - architecture chitchat' threads ever been entirely on-topic? Ever since the Bulldozer one they seem to have become a de-facto 'microprocessor chitchat that doesn't necessitate its own thread' thread! :P It's nice to discuss stuff which is a bit off-topic without starting a hundred new threads which would likely get no or few replies.
I think having AMD in the title just keeps the riff-raff out ;) :D
Point taken though! I was wandering a bit far, just thought the Exynos uArch bit might have been of interest, and semi-relevant.
Back on topic - if anyone hasn't seen it yet, Anandtech have a decent GloFo interview up: https://www.anandtech.com/show/12438...lobalfoundries
It includes things like the motivation behind skipping the 10nm node, where FD-SOI nodes are going, details on 12nm and also some more about EUV plans.
Definitely interesting. Some pulls from my PoV:
So 12nm Ryzen should clock to around 4.4GHz. If they get much in the way of IPC improvements too they'll be knocking on Intel's stock-clocked door for ST workloads...Quote:
Customers want performance kickers as well as some density improvements like in 12LP, which offers about a 10% performance improvement and a 15% circuit density improvement.
Am I reading that right? Moving from 14nm to 7nm the same logic will be almost 1/3 of the size?!Quote:
... For example, we have a shrink on our 7nm from 14nm that is 0.37x scale. So it's more than 50% scale at a logic library level. ...
*snerk* ;)Quote:
At least one of our competitors spends a lot of time talking about the pitches and the things like that ...
The FD-SOI bit was interesting - diverging their performance and low-power lines to give customers more options. I wonder if we'll see AMD looking at the FDX processes for any of their embedded IP, or if their plans are all too performance-oriented for that to be worthwhile...
I think the memory latency issue is probably something I hope that can get improved with Ryzen+ as a Core i5 8400 shouldn't nearly be as far ahead in gaming,than it should be! Look at say a Ryzen 5 1600X against a Broadwell Core i7 6800K. In non-gaming stuff its very close clock to clock,but in gaming it does seem significantly behind Broadwell overall.
At last i have a better understanding of the differences between LP and LPP, at least i think i do as i suspect the FinFET with 2 fins are the low power ones and the 4 fin is the performance plus.
:clapping:Quote:
one that is a 2 fin library, so it is ultra-dense, and another that is a 4 fin library for maximum performance.
Hey all, recieved and used my boot kit for the 2400g. Updated my ASrock A320m Pro4 up to version 4.50, removed the toy APU and put my 2400 back, built it *just so* and now find I still can't boot. Any suggestions? And yes, I know this could easily go in tech issues thread.. but it's Ryzen.. in Ryzen thread.. and you love going off topic... and I get an audience.. :hexlub:
Specs:
ASRock A320M Pro4
Ryzen 2400G
2x 4GB Team Group Vulcan 3200MHz CL16, slots A2 B2
Crucial MX300 525GB M.2 SSD
Superflower Platinum 450w PSU
rando ASUS optical drive.
Currently, on power on the system will sit and think for what seems to be forever. If I power it off, then on again, it will begin looping an on/off cycle. Never displays anything. Of note, it worked just fine with boot kit APU, and I have it connected to a VGA only display (which was fine for that earlier).
On a side note regarding the phones,I was kind of surprised by the quality of the audio output from the Huawei phone I had(it sounded good even though really high end headphones costing like £2000 to £3000),and it appears they integrated a proper DSP:
http://askmeblogs.com/huawei-p9-lite-review-on-a-diet/
https://www.gsmarena.com/hauwei_p9_l...iew-1461p7.php
It makes me wonder whether a cheapo phone actually has better sound quality than most of these high end motherboards with all their "special audio sections". I do miss Nvidia SoundStorm TBH!!
Try booting with no drives connected,and one stick of RAM. Have you got any other DDR4 you could try??
Edit!!
https://www.techspot.com/article/158...-motherboards/
Quote:
What's interesting is that while the Asrock A320M lists up to DDR4-3200 memory support via overclocking, I wasn't able to get either Raven Ridge APU working at this speed. In fact, 3066 failed and I even ran into a few stability issues at 2933, forcing me down to 2666 and thankfully the system was completely stable here.
In fairness to A320 boards, I got my hands on the most expensive Asrock A320 model, the A320M Pro4, which makes little sense at $65 but I wanted to see if memory support was any better. Here I was unable to post with the 3200 XMP profile enabled and was forced to run at 2933, which did appear to be stable. I should note that increasing the DRAM voltage didn't help on either board.
Now I'm not saying all A320 motherboards won't work above spec for the memory but I tried two different models and neither worked. I suspect the same might be true for the ultra cheap $60 B350 boards that share the same physical design as the budget A320 models.
Eh, while I don't like the concept of stealing RAM from my primary to get it booting, will happily do that if ASRock release a BIOS revision for RR which actually works! Suppose it could be tested fairly quickly now and prove / disprove the issue.
Can easily remove the M.2 but I can't see a blank M.2 doing the system much harm, especially since it was fine with boot kit. Just searching for reasons that it might refuse to boot for now.
Yep, I'd seen it, and I have to assume that ASrock has a BIOS patch on the way to get the board at least to its proper 2933MHz rating, which is what I meant. There's 2666MHz rated DDR4 in my primary PC but obviously I'd have to gut it to check if RAM speed is the problem (and not know how long for either). Swapping kits would put me at half memory capacity for now, might not be worth it until there's a new BIOS version for the A320M.
Tried it out with 2666MHz CL15 single 8GB stick. Removed the M.2 and disconnected the CD drive (heh) but to no avail. It's like I never flashed it to 4.50, though I can guarantee that it is version 4.50. Still power cycles before loading BIOS.
AMD CPU and GPU marketshare has seen some decent increases:
http://www.tomshardware.com/news/amd...are,36592.html
Leaked Zen+ and Zen scores compared to measure IPC(taken off Reddit):
https://imgur.com/a/Do2dd
A Korean website tested an NUC with the integrated Vega graphics:
https://translate.google.co.uk/trans...%2F&edit-text=
https://translate.google.co.uk/trans...%23&edit-text=
GPU seems to be in-between a GTX1050TI and a GTX1060 in performance.
Dell also CBA making more than a few AMD systems it appears,as Intel is bigger:
http://www.channelpro.co.uk/news/107...ucts-says-dell
Worth noting that the interviewee here is the CTO of Dell EMC, rather than Dell Inc.. Given EMC is an enterprise storage and virtualisation company (albeit a wholly owned subsidiary of Dell) his worldview is likely to be somewhat different to the people in charge of Dell's consumer and workstation line-up...
It hasn't stopped the former Intel chief engineer going out and using it to spread fud about AMD on Twitter,and didn't we see the same lame excuses back in the day?? Isn't that the same area Intel has identified as a "growth area!" ?? ;)
It seems Baidu and MS don't agree with him,since they are early adopters of Epyc,and so is QNAP for their usage of Zen based products. Guess what these companies use them for?? ;)
Edit!!
I suspect its more the case Zen based products would end up being cheaper than comparable Intel ones,and that would force Dell to drop pricing on the Intel based ones,and that would probably annoy Intel as they would have to drop pricing,and if they didn't Dell would have to drop margins.
https://pics.computerbase.de/8/1/9/8/6/9-1260.jpg
Looks like the new Intel control panel is basically a blue version of the AMD one!! :D
Edit!!
Some benchmarks of the Acer Nitro 5 which has an RX560:
https://www.notebookcheck.net/First-....277336.0.html
It looks like the chassis has enough cooling for the Ryzen 7 2700U in the laptop to run at a 25W TDP.
Yeah,its quite funny!! :p
Edit!!
Ryzen 7 2700X clockspeeds apparently leaked:
https://translate.google.co.uk/trans...%2F&edit-text=
Runs at 3.7GHZ~4.1GHZ it seems.
The first Ryzen 3 based laptop has been released:
http://www.tomshardware.com/news/del...ega,36612.html
Ryzen 3 2200U goes down to 2C/4T and only 3 Vega CUs. Unless yields are very bad, they must be throwing so much potential away there.
Two dies to rule them all for now, but if they get their volume up making a much smaller die would be essential. Getting over 200mm² down to under 100mm² should be possible with half the dies and 3/10 of the CUs.
Without fully understanding AMD's financials it's hard to say for sure, but it's not as simple as just chucking out an arbitrary number of die variants for each market niche. Exacerbated by current cutting-edge nodes, mask costs can have a substantial impact to margins - the target market needs to be large enough to offset the initial costs whereas conservative binning is very flexible and allows greater selectivity of the higher-end parts. In addition, yet another die variant adds complexity to inventory management, but again binning is flexible and allows fast response to market demands vs lead time measured in months for wafer orders.
It's not necessarily correct to attribute segmentation/binning to yield either - Intel have long been heavily stripping features from cheaper CPUs which have little to no impact on yield e.g. SMT, instruction sets, cache sizes, etc. Same for their chipsets which generally all use an identical die. Intel still had plenty of market for dual core dies though given their reluctancy to allow 4 cores in 'mainstream' products (at least until competition forced their hand) and selling dual cores as 'high end' in the laptop market.
Stripping half of the cores and most of the CUs wouldn't necessarily save as much as 50% of the die anyway - take a look at the die shot.
Okay, that's a fair point.
https://i.imgur.com/OZvyjfo.jpg
So the one CCX takes up ~19.5% of the die, and all the Vega cores take up 31.3%.
Meaning half the die is taken up by fixed stuff. Half the CCX would only save 9.7%, and 3/10 CUs about 21.9% for a saving of around 32%. At around 210mm² for Raven Ridge, that would make such a die about 140mm² still a lot bigger than the Sky/Kaby/CoffeeLake 2C dies which are around 99mm².
I'm sure once AMD's volumes are high enough it might eventually be worth their while to make smaller dies - although then again if they want to hit, for example, a 5W target they might be better off with 6CUs running slower etc. That is, they have plenty of options.
I could be wrong but I suspect AMD are aiming more to push quad cores into mobile rather than their dual cores so they might not want 2C volumes to get that high. However maybe if they also use a cut-down die in their embedded products it could make more financial sense?
There are loads of BR based laptops still in the channel so there is little incentive for AMD to want to price 2C/4T Ryzen too low ATM.
I think the key is product positioning. If the 2200U is positioned that close to the 2500U ($729 for a 2200U/12GB/1TV, $899 for a 2500U/16GB/2TB), where does the 2300U slot in? Current pricing makes it look like the 2200U is directly replacing the A12 9700P (e.g. an Inspiron 15 5000 with A12/12GB/1TB is currently $699 after a $100 discount) - it just seems really odd to replace the top SKUs from Bristol Ridge with the bottom SKUs of Ryzen. I'd've expected the 2300U to take that slot, and the 2200U to replace the dual core A9 etc. If the 2200U is replacing the quad core BRs, are there going to be any sub-£500 Ryzen laptops? Or are they just going to bag off that entire market? Seems like an odd decision, unless there is going to be a lower spec range of processors to fill out the bottom end (maybe that 2C/3CU single channel die is out there after all!)...
If they don't hurry up and get something in the market, though, they're going to lose my money. This old HP is very much on its way out, and if there's nothing AMD in my price range when I need to replace it....
The problem is you can get a KL 2C/4T CPU and a 940MX DDR3 for £340:
https://www.ebuyer.com/810641-asus-a...a556uq-xo1470t
Sure it only has 4GB DDR4 and a 1TB disk,but still!!
Edit!!
Model with a 1080p screen and an SSD for under £500:
https://www.ebuyer.com/805830-asus-a...a556uq-dm1319t
Second Edit!!
You can get laptops with MX150,GTX950M and GTX1050 for between £600 to £650 too. Sure they are all dual core,but Ryzen Mobile pricing does seem rather high TBH.
I'd already spotted that one CAT - it's why I'm getting frustrated with the way AMD's handling the mobile Ryzen launch. Laptops are a bigger market than desktops, they've got a great range of laptop processors, but they're trickling to market at ridiculous prices. They should have a good range of 2C/4T laptops in the market at ~ £500 with a 1080p screen.
That ASUS has a spare SODIMM slot to, so for an extra £40 you can bump it up to 8GB. Hexus very kindly compared an i5 7200U + 940MX against Ryzen 5 in gaming. Let's see how that went: http://hexus.net/tech/reviews/laptop...wift-3/?page=7
Yeah, not great for Ryzen. I'm this <-> close to just buying that ASUS now....
It does have a 940MX/MX130 GDDR5 in the Hexus review,but the cheaper Asus has the MX130 GDDR3,so a bit slower,but still for under £500 its not too bad.
I do have some concerns about AMD driver support too - if you read through the comments thread of the review. The drivers on the Acer site are much older than the official AMD desktop RR APU drivers,but you can't download any drivers for the mobile APUs on the AMD page(!),but you can for both versions of RR,ie,desktop and laptop. Also as I mentioned one YT site forced Vega desktop drivers and it worked fine and performance went up a decent amount.
Even the desktop RR launch was not ideal with half the AMD OEMs only shipping RR ready BIOSes,the week before the reviews hit.
Plus you can get laptops with the MX150 and GTX1050 for even cheaper:
https://www.laptopsdirect.co.uk/-9s7...68/version.asp
https://www.laptopsdirect.co.uk/refu...ea/version.asp
Both are 4C/4T.
GTX950M DDR3 laptop for £500 with 1080p screen,SSD and HDD:
https://www.saveonlaptops.co.uk/NX.G...G_2186600.html
Similar with one RX460 GDDR5 and FX CPU:
https://www.saveonlaptops.co.uk/K550...T_2108500.html
One with 4C/4T KB CPU,MX150 GDDR5 for £590:
https://www.saveonlaptops.co.uk/9S7-...r_2197296.html
Then this really good deal for £700ish:
https://www.saveonlaptops.co.uk/ZX55...T_2145527.html
Edit!!Quote:
ASUS Gaming ZX553VD-DM969T
Intel Core i7-7700HQ Quad Core Processor
15.6" Full HD Screen
Microsoft Windows 10 Home 64-bit
8GB DDR4 RAM
1000GB HDD
No Optical Drive
Dedicated GeForce GTX 1050 2GB Graphics
USB3 | HDMI | Bluetooth |
Red Backlit Keyboard
ZX553VD-DM969T
Repeated the MSI PL62 twice.Oops!! :p
I had a look on saveonlaptops,and you can get that Asus ROG laptop with a 128GB SSD an 1TB drive off the shelf for £750. Not sure if the Acer Spin 3 makes much sense at £700,especially after looking at this review of the Asus laptop:
https://www.notebookcheck.net/Asus-R....200646.0.html
The only advantages the Acer Spin 3 has are a longer warranty and potentially better battery life under heavy combined CPU and graphics load. But a GTX1050 GDDR5 is much faster than a Vega8 IGP with dual channel RAM. Even the £700 model has a spare M2 slot so you could get an SSD and install one yourself.
Edit!!
Even the HDD is a 7200RPM unit which is quicker than normal as many laptop HDDs tend to be 5400RPM to 5900RPM units.
Second Edit!!
That mobile FX and a RX460 GDDR5 is faster than a 2C/4T KL CPU and an MX150 GDDR5 in the games tested.LOL.
Laptop pricing is all over the place at the minute!
The ASUS with RX 460 is kind of tempting ... but the 1366x768 screen just might be a deal breaker, and the lack of DVD drive counts against it too. Notebookcheck's review isn't exactly glowing! Sadly it's just the wrong balance for me, for all the good price:performance ratio.
The F15 is also kind of tempting, but I'm not sure if the i3 is a compromise too far to get the 8GB RAM and 950M over the 940MX - particularly when it's the DDR3 950M, which iirc really sucks performance out of it (there's a more recent F15 in the house already with i5 7200U and GDDR5 950M which is a lovely machine).
*sigh*
The sad thing is that I really don't need anything my existing laptop hasn't got. A 1080p screen is the only meaningful upgrade to me, but getting 8GB and 1080p with any processor seems to push you right up to £500 already, and I kind of feel that if I'm going to spend that much I ought to get at least some CPU and GPU upgrade as well....
£430 will get me an i3 7130u with 8GB/1TB/1080p. Not sure the HD620 would be an improvement graphically on my APU though. £500 gets me the Acer F15 you linked, or a Lenovo with A12 9720p (also 8GB/1TB/1080p). The F15 at least has a (small) SSD, but is £500 too much for a DDR3 GPU? ARGH! Just can't decide ;)
How about something leftfield: https://www.amazon.co.uk/HoMei-Dedic...dp/B079H1BH6R/ - it uses the desktop Pentium Gold G4560 and has a 2GB GDDR5 MX150 (plus the standard 8GB/1TB/1080p) for £570. Or for the same price there's an MSI with an i5 HQ, and IPS display...
The PL62 has a 4C CPU,and a GTX1050 2GB GDDR5 for £580 from Ebuyer. If you are spending the better part of £500ish,you might as well get at least a reasonable card.
Also there is a 4C version of the Asus ROG laptop:
https://www.saveonlaptops.co.uk/ZX55...T_2145526.html
It has a 7200RPM drive and a backlight keyboard for £650.
Edit!!
You can get an Atom based laptop with a 1080p display for £130:
https://forums.hexus.net/retail-ther...-129-99-a.html
So,it makes me wonder how expensive having a 1080p screen is!!
Problem is, I don't *want* to spend the better part of £500. I want to spend as close to £400 as possible, but 8GB and 1080p are non-negotiable, and "better" CPU and Graphics performance should be a given.
I've just done a bit of comparison of my Fire Strike scores to an A10 9620p (which is the same 4C/6CU configuration as my A10 4600m), and while it's not necessarily indicative of overall performance, it seems to be about 25% faster at everything. I guess that's not bad. esp with the lower TDP.
OTOH, the cheapest 8GB/1080p laptop I can find with the chip is a 17" HP at £467* - so not a great improvement on what I've already been looking at! At the same price I can get a similar 15" with an i5 8250u...
*EDIT: that's actually with a 10% off promotion on all HP laptops, on their website. Brings the Ryzen Envy 360 down to £719...
Its not a bad price for a Core i5 8250u based laptop - but shows you how massively overpriced the Ryzen 5 2500U is.
The issue with a £700 Ryzen laptop is you can get an Asus ROG laptop with a Core i7 7700HQ and a GTX1050 and a backlit keyboard for £700ish. At this point Ryzen is just not really competitive currently. I would rather have the Asus laptop as it just seems a better device than the HP Envy which has a rather average screen apparently,and the CPU and GPU will blow the Ryzen 5 2500U out of the water,and the laptop is probably just built better and with a better screen. Remember,its reduced down to that price by £150 to £200 IIRC.
Edit!!
TBF,you can get low clocked Core i3 laptops with a 1080p screen for as low as £350,but the low clockspeed means even that AMD FX-9830P based laptop will have better performance,and it has a 1080p screen(not a 768p one).
Second Edit!!
No it doesn't,oh well! :p
Loads of deals on the Dell refurb store:
http://outlet.euro.dell.com/Online/I...WiUHjTxyv70%3d
1080p displays and the Core i5 8250U for £350ish. Core i5 8250U,SSD,HDD and a 940MX GDDR5 with a 1080p display for around £500ish.
Edit!!
Another laptop but from Argos but with a GTX1050:
https://www.ebay.co.uk/itm/Medion-P6...UAAOSwehVaknWr
If that was the case Dell wouldn't be offering the 2C 2200U in a $679 laptop - they'd be offering the 4C/4T 2300U instead. That Dell could potentially be a massive own-goal for AMD - people will buy it because they've heard mobile Ryzen is gaming-capable, and suddenly they've spent almost $700 on a dual-core, 192-shader laptop that isn't much better than Intel HD graphics in games. It's the space the 2300U should be in, with the 2200U in the low end parts. Otherwise the 2300U is basically pointless...
Whats the point of mobile Ryzen 5 if Core i5 8250U laptops can be had cheaper and laptops with a Core i7 7700HQ and a GTX1050 can be had for similar money?? Its all a bit pointless then. There is no excuse when the AMD is a single chip solution and the competing ones use two to three chips and needs a more complex cooling solution too!!
I suspect that we're seeing a bit of an early adopter tax still, and some of the cheaper Intel laptops are heavily discounted, which simply won't happen with "new" Raven Ridge designs, so I don't think we're seeing a clear picture of the landscape currently. At RRP most of the Ryzen 5 laptops look to be about the same cost as the equivalent 8th gen i5, which tbh is what I'd expect given the comparative performance.
What I find most odd about the laptop pricing is that the desktop pricing has been totally on point. It could be that the Dell 2200U is an anomaly and that most OEMs will be pricing their 2200U laptops more sensibly, but it's such a jumble I really don't know what to do atm.
I've been doing a bit of cross comparison (to my A10 4600m, and to the leaked 3DM11 scores), and it looks likes - graphically - HD620 is a bit faster than my A10, and a bit slower than a 2200U. Does make me think that I might as well just get something cheap and Intelish. I'm quite liking the look of the 13.3" Lenovo 510s, and I can get it for < £400 with a 1080p screen and 4GB. I've dug the service manual out and while upgrading the RAM (and possibly the SSD) isn't entirely straightforward it's well within my capabilities...
Started a new job yesterday, spent most of the day trying to beat my Dell Precision laptop into some shape usable for coding. It is one of their range that can be delivered with Ubuntu on it, but I want Fedora so it was delivered with Windows so it has the license if ever needed. So I have to fight the storage controller settings, the switchable Intel/Quadro graphics and the Thunderbolt docking station.
I would have loved to have had a single AMD chip that just works. For now I have given up on multi-monitor support. I doubt the 7700HQ is life changingly faster than a mobile Ryzen at compiling code. I am starting to wonder if I should have gone for that monster Asus gaming laptop with the Ryzen 1600 and RX 580 in it. Lack of portability put me off on that one, the 1 hour battery life would probably never hurt me.
You won't get a 1050 in a Precision laptop, this puppy has a Quadro in it. Utter waste of money as Optimus graphics switching means I still have to use the Intel graphics drivers.
I know Phoronix had problems with RR under Linux, but AMD's recent record on Linux drivers is pretty good so I'm sure it will get sorted soon. So far this laptop has been a real source of stress, I might get it sorted in the end but this is the established technology from the main players of Intel/Nvidia/Dell making a business laptop that is Ubuntu certified and I am still struggling, and I am sure this is as good as it is going to get.
Its probably even less of a chance of getting an AMD card in there!! :p But for the average consumer,almost all of the RR laptops seem to be sub-optimal in some way or another,and way too expensive,and under Windows AMD can't even support the laptop RR,and when people have forced desktop Vega 56/64 drivers on it and there was performance jumps,one has to go LOLWTFBBQ at it. I mean the latest RR laptop drivers are months out of date and you cannot even download them on their website(BR you could). By the time they sort it out,Intel would have also sorted out cheaper CFL SKUs.
It looks like HP has leaked 68 different configurations of AMD Ryzen powered Elitebooks:
http://www.tomshardware.com/news/amd...ook,36613.html
Hopefully these will be better than the currently launched ones.
The nice bit about AMD graphics under Linux is that you can just install the latest OS and because everything is open source there is no hunting for drivers as the entire stack down to Vulkan is included. My home desktop has an R9 380 because that was the first chip that really got that support, and I really don't want to go back to downloaded drivers.
The dev target is an Altera based system, more so Intel downloads trying to get a dev kit that even installs.
Hmmm, 13.3", Ryzen 3 2300U, 256Gb SSD, 8GB of RAM. What do we reckon that'll cost, north of £700? *sigh* :rolleyes:
EDIT: oh boy, just looked up existing HP Elitebook prices. If you buy direct from HP you can get the current gen 745 (G4, 14" FHD) with an A10 9800B, 8GB and a 256GB SSD for £790. Intel versions (Elitebook 800 series) start at over £1000, afaict. Given the tendency to price Ryzen laptops like their Intel counterparts, I think I know what to expect from the 700 G5 series... :o
A purported leak of some Zen+ scores:
https://videocardz.com/75185/first-b...ve-been-leaked
Memory and cache latencies seem a decent amount lower!!
I've noticed something: the efficiency of the RR vega parts is unusually low.
Looking at historic AMD cards, they tend to score lower in time spy than similar performing nvidia GPUs - example one (570 behind the 1060 by ~8%, but equals it in games) and example two (V56 beaten by 1070ti by ~10%, and about equal in games). There's something odd about the RR scores, in this light - vega8 is well ahead of the 940MX (by 11%), which ought to correspond to a performance lead of ~20% (assuming the 3dmark:fps ratio stays constant), but the 940MX instead beats vega8. This is a major discrepancy - 20% more than an MX130 is getting close to MX150 performance, and so it looks to me like the outdated drivers are massively hobbling the RR APUs. Typical AMD launch!
Ryzen+ line-up leaked:
https://videocardz.com/75194/amd-ryz...ormance-leaked
That is a big increase in clockspeeds!!
Edit!!
It could be a fake as the NDA date says 2017!!
I wonder what's happened to the 2800(X)? Maybe they're waiting on yield ramp for higher clocks?
I'd think about putting money on it being fake. The slide looks a little too familiar. Plus why on earth would the slide detailing the Ryzen 5 2600 also have a Ryzen 5 1600 on it?! I'm also not buying a 105W TDP on the 2700X when all other AM4 processors have been max 95W. That'd be at serious risk of breaking backwards compatibility unless they'd over-specced the platform originally...
Also looking at some of the results they looked off - FO4 and Skyrim show big differences but use the same engine,and also the uncapped FPS difference between the 1800X and 2700X is only around 10% for FO4 and makes it equal to a 8700K(which is a load of Moose droppings),and then the uncapped Skyrim are 40% higher on the 8700K using the same engine.
If it was an AMD slide,then its downright deceitful:
https://cdn.sweclockers.com/artikel/...dd39f1da29c8fa
https://cdn.alza.cz/Foto/ImgGalery/I...ut-4-1080p.Png
https://tpucdn.com/reviews/AMD/Ryzen..._1920_1080.png
An 1800X is NOT 10% slower than a 8700K at 1080p using a GTX1080 in FO4 especially if you look at minimums,unless they found an area which was mostly GPU limited.
If that leak is to be believed the prices (or at least RRPs) seem a little marked up. Good speeds though, should be ok. Rumours GloFo CTO says Ryzen 3 @ 7nm will be 5ghz ish.
Another leak which apparently confirms the line-up:
https://videocardz.com/75217/amd-ryz...eer-and-renoir
Interestingly that implies that a) the 1500X and 1300X are being dropped (silently?) from the CPU line up this quarter, and b) that there won't be any Gen 2 quad core CPUs at all - it's 6/8 core CPUs and quad core APUs. It's a very simple looking product stack, although the disparity in numbering between mobile (2700, 2500, 2300, 2200) and desktop (2700,2600,2400,2200) looks rather Intel-esque in its complexity...
Another £500 Core i5 8250U laptop with a 1080p screen:
https://www.box.co.uk/N24T6-Dell-Ins...&siteid=116019
Seems to have this graphics card:
https://www.notebookcheck.net/AMD-Ra....214534.0.html
The ones on the Dell site have 4GB of GDDR5 but this model has 2GB of VRAM. Amazon indicates it is 2GB of GDDR5.
Edit!!
GPU performance will be between a 920MX and 940MX/MX130 DDR3:
https://www.youtube.com/watch?v=y7P8xN9pqLM
So probably quicker than the Intel IGP,but slower than the Vega8 IGP I suspect!!
Second Edit!!
AMD needs to get those drivers out,as if at least if graphics performance increases it might fend off some of these cheaper Core i5 8250U based laptops. It does make me wonder whether its not helping with battery life either??!!
We did see high-end ryzen launched first with the 1X00 series
If the model numbers are true especially as AMD on Twitter actually shared some of them briefly,I do wonder whether AMD is going to launch a higher binned Ryzen 7 2800X with a higher TDP like 125W?? If the X470 has lower power consumption that might negate to a degree any of the CPU power consumption increases.
Having said that upto a 400MHZ increase at 65W TDP and upto a 250MHZ increase at 95W TDP is not actually bad after only one year IMHO.
My takeaway from the leaks - if they're true - is that we should see OCs at around 4.2GHz - 4.3GHz. 6C/12T @ 4.2GHz for $199? Not too shabby....
As to the line up; note that the SKU transition slide very clearly shows the 2200G replacing both the 1200 and 1300X, and the 2600 replacing the 1500X - the 2600X covers both the 1600 and 1600X, and the 2700X covers both the 1700X and 1800X. Sure, it leaves some room for a 2500 and 2300 to slide into the stack, but is there really enough RRP and performance room for them? The 2500X would have to be priced between $169 and $199 (unless they drop the price on the 2400G when they introduce it)...
Depends what you mean by nicer, I suspect. AFAICT the latency is still determined by the fabric rather than the IMC, and I've seen nothing to suggest that the fabric speed is being uncoupled from the RAM clock, so it's still going to want nice fast RAM to get the best performance - the latency improvements in desktop Raven Ridge latency are better than you'd expect from clock speed alone, but they're getting a good 10% boost just from using 2933MHz RAM v 2666MHz for Summit Ridge.
I'd assumed the purported leak on videocardz was running both chips at their rated RAM frequencies (so 2.4 GHz vs 2.9 GHz for the new chips), but they're both running 2.7 GHz and the new chip still gets an 11% drop in latency
Crikey:
https://m.hardocp.com/article/2018/0...onsumer_choice
Is Nvidia that worried about AMD Ryzen and the AMD-Intel partnership??