They seem impressive, the 1800X looks like it beats out the 6800K in multi-thread cinebench 1601 Vs 1474 multi-thread Ties with the 6800K @ 162 in single thread cinebench. @ $499 vs $1050 for the 6800K. I'll wait till, the reviewers get thier hands on them and put them through the paces. Currently, looks like a 1800X will be replacing my 3770K @ 4.6 in the near future.
Looks like the 1800X is already sold out on preorders would I take a huge lost, if I go with the 1700X over the 1800X? Should I wait? I keep hearing they are the same chips, but the 1800X is clocked higher and binned. In theory, if I throw a aio or some kind of water cooler on it and let XFR do its thing.. or manually oc'ing.. Is it possible to get atleast 4.0 out of it? 1700X chip, bundles with Asus Mobo's are going for $569 on Amazon.
I think they are eventually but they are releasing all of the desktop stuff first if I remember correctly.
R7 1800X beats 6900K (not 6800K) also Linus showed that AMD put Intel in better or even PCs for all tests compared to AMD PCs so RYZEN may be just a little bit better. vid: ^here he personally does all AMD tests himself, showing test PCs and some info that you may missed About OC no official info about it, well actually there is one: it OCs well on LN2 and reached best 8c/16t score in Cinderbench - so it's fastest 8c/16t CPU stock and fastest 8c/16t CPU after OC - YET vid: sorry couldn't find original ^this one is copied IPC is still lower than Skylake/Kabylake (still it exceeded expectations, nobody expected RYZEN to be as good as it is) but AMD CPUs clocks better and have lower TDP too, also way better price/performence ratio. Personally I'm more interested in R3s will they clock better than i5s too? Because it looks like Zen require a lot of voltage over "X GHz" point and may not OC well on air and water and it's IPC is still not as good as Intel's, so they may win in multi-core apps but still lose in games that relay on single/dual core performance. The difference between 1700X and 1800X after OC is unknown at this moment. Nobody can answer it. Anyway do not preorder anything (I'm AMD guy, but literally pre-ordering things is risky and things may not work as good as it looks) I'd wait for 2nd March and watch benchmarks from other apps and games, OC etc. There will be all info about 1700, 1700X and 1800X available, as some redactions already have production CPUs and are doing test right now to release all data when NDA ends @ 2nd march. Yes, Raven Ridge (new AMD APU lineup) is going to be released in 2nd half of 2017
Ah, while Raven Ridge might be around the corner, it's important to remember that Bristol Ridge is already here for OEM's, and later for the consumer. This is to hold APU users over until Raven Ridge is finished. Unfortunately... AMD just refuses to let Bulldozer die, so guess what architecture the first AM4 APUs are using. That's right! Bloody Excavator. >_>
So these chips* don't have any integrated graphics capability? So if your graphics card takes a dive, you're SOL until you can get a replacement?
These are not cards, they are cpu's and not having integrated graphics isn't a new thing, Intel has been doing it for a while with different sku's. Also a gpu doesn't just randomly die, it will give you warning signs of its imminent failure i.e. bluescreens, artifacting, crashing, instability. You will have plenty of time to get a replacement, and most cards will either fail withing the return period or way after the manufacturer warranty is up.
Yeah I know, I mean't chips. I was just curious if it would run without a graphics card at all. I'm holding off on building my new PC for these CPUs. One of these backed by a RX480 sounds like a nice combination.
@Drivver Thanks, for the correction. I'm also alttile iffy on preordering and it "not turning out". This will be replacing my 3770K @4.6 GHz. I did some single thread testing with cinebench last night and it looks like my 3700K. Holds 163-164 CB
Single core performance is what must be checked out, BeamNG really needs good single core performance, at least in 0.8, not sure if that changes in 0.9, but as always, never jump into new thing, wait 3-4 months after release and you probably have much better information to go by as well as lot less premium to pay for. Also, comparing new upcoming AMD CPU to old intel CPU might give bit misleading truth, Kaby Lake single core performance might be one which to compare at least so that one can see where AMD actually sits at. Workstation platform of Intel is soon replaced and again 6800k and such might be close to AMD prices used, is there point to get such used depends from AMD's single core performance I guess. Multicore performance is nice and helps with many vehicles for sure, but currently single core performance is what seems to be one where things have not progressed much between generations of CPUs, that seems to be holding back performance with most of the times on my computer and if AMD is slower than Intel on that, then I would not be jumping to that ship, multi core performance seems to be quite impressive, but I would wait 4 months and see truth to be spread open before making any decisions.
Single core performance is similar and even better in some of these CPUs vs the Intel models according to the single core Cinebench results.
Well whats your definition of "practical" applications? A cinebench score tells me exactly how well it will perform/compare in cpu based rendering software like cinema 4d; my 3d modeling/rendering application of choice, so it is very practical for me. Everybody's idea of what is practical is different, so it is completely dependent on the end user to decide if these products are practical or not for their workflow/gaming needs.
Skylake = Kabylake, Kabylake is just more mature 14nm process of Skylake so it clocks just a bit higher with same TDP, same IPC. From some reliable leaks it looks like RYZEN OCs fine, 1700 to 4GHz on good mobo and stock 1800X is inferior comapred to it. http://www.tweaktown.com/news/56422/amd-ryzen-7-1700-4ghz-1800x-performance-329/index.html I wouldn't say single core performance is better for BeamNG, my friend's laptop with i5 4220H or somehing like this handles BeamNG on high settings and is limited by GPU (GTX 960M), and this CPU clocks at ~2.8GHz. My old PC with i5 4690K @ 4.5GHz, 8GB RAM and very high OCed GTX 970 was able to run 15 covets driven by AI on utah at medium settings with over 20fps, up to 7 covets it was over 60fps, then changing to G3258 @ 4.3GHz (same PC, cooler, ram GPU etc) and it handled 3 covets over 60fps and just 7 covets over 20fps (look at this numbers. Amount of cores 2x covets -1 = over 60fps, then 4times amount of cores -1 = over 20fps, but just 4x more covets than amount of cores and you got like 10 fps. So 1 haswell core at over 4.3GHz is fully capable of running 2 covets over 60fps, but 1st core also have to deal with some graphics actions, world physics simulation so 1st core that have to deal with BeamNG world is limited to 1 vehicle, then next "free" cores can handle 2 vehicles and still keep 60fps because they don't deal with as much operations as 1st core) from this test I can tell you that best choice for BeamNG is multicore CPU with high IPC and clock speeds - R7s have best ratio to cores, IPC and clock speeds so it may be great for BeamNG. Yeah, yeah very high single core performance counts on not properly made maps, due to T3D roots, butvanilla maps doesn't impact CPU performance that much, so 1st core isn't bottlenecking with map and single vechicle running on it, and then more cores = more cars (ofc there are some GPU limitations too, as vechicles are very GPU intensive too) Forgive me messy explaination and errors I'm very weak atm, I hope you understood what I meant. Other than cinebench there are also handbrake and multitasking benchmarkds done. But Just wait till 2nd march and EVERYTHING will be clear. IMO all OCed RYZENs will work great with BeamNG - we'll see soon if I'm right or not
In vanilla maps single core performance is important if you drive at night, shadows use CPU quite a lot and multiple lights + shadows means that graphics thread gets lot of computing, so much that with single car I'm nearly at limit at East Coast USA town with i7-6700. Of course that might change in a future. I'm leaning towards waiting KabyLake-X, Skylake-X or whatever those upcoming Intel workstation CPUs are called, but it is not before summer or maybe not even then that there is reliable information about those. Certainly AMD would give lot performance for the money, but I think that this year is good in many ways for pc builds. I probably will go for used x99 platform, however I'm not really fan of anything, maybe little bit BeamNG fan, but it really does not matter if I have 3 or 5 letter brand CPU in my computer, performance / noise (heat) and cost is more important and with my current system I can wait easily until next year if needed to see if there is really some good single core performance boost coming to my reach, BeamNG is major thing where I could use more CPU power, but even web browser could use little bit of boost.
I have only DDR4-2666, but what benchmarks I saw there was quite bit of performance difference between 2133 and 2666 with i3-6100, might be that currently with higher end CPUs there is gains available with faster ram, back when I checked those benchmarks it was 2800-3000 where gains started to be quite small, but that might indeed been improved already. With BeamNG using all cores and quite bit of ram too, I would be easily lead to believe that BeamNG is one title to gain from faster ram. With K CPU it would be quite simple thing to test out I guess. Not sure how much AMD would lose with slower ram though, it is something that needs to be seen in a future I think. I hope that 0.9 has Banana bench fixed and some comparable testing could be made.
Going of of the cinebench single core score that AMD has shown for comparison My 37770K @ 4.6 GHz 163 1800X @ 4.0 Ghz 162 So it takes 600 MHz more for my 3770K to produce the same or similar single core score. If I were to switch over to ryzen, (going by those cinebench scores at stock 1800X speeds) in theory I will be maintaining basically the same single core performance. but I will have more cores and threads to play with BNG, and it at will be doing it at slower clock speed. Is this correct? I wanted to preorder a 1800X and be the first one to put it under load via banana bench to see if this is true. But ill wait and see for the reviews. Once I decide on ordering, it will likely be cooled by a AIO and ill push it as far as possible clock speed wise.
