AMD's Ryzen CPUs (Ryzen/TR/Epyc) & Vega/Polaris/Navi GPUs

You don't have to read it word for word...

But from the second link above; when Intel is hitting ~36.67ns to ~43.08ns and AMD is in the ~111.5ns to 224ns range - stating Intel is world class is not a stretch - regardless how many competitors are in this field.

What is ridiculous to me is that AMD blames the tools used to measure this... why don't they just give us their tool(s) with source code so we can measure accurately ourselves? (Even then; they're still much slower than Intel's caches...).

 

My point is, on single thread performance it does help the 99 task or better of a hundred complete in 0.5 seconds instead of 0.45 but that is not what I want a threadripper for. It is for the one or two tasks per week I now have the machine do for 3-6 hours encoding and backing up video library. Now those tasks can handle all the threads I can throw at them, so if I can reduce this to 1-2 hours I am so there.

As far as cache, you discount a few things.

1.) cache latency decrease is not in direct correlation to performance increase. It contributes somewhat but it is again not directly proportionate as can be seen.
2.) size matters, in this with Ryzen having a larger cache per CCX data and/or instructions are less likely to be thrown from cache. Saving any information from being reloaded is a time saver.

I am sure there are some other tips and tricks we are not privy too as well.

 

When a processor is waiting to be fed instructions/data: cache latency is very important. It isn't the latency that is not a direct correlation - it is the cache misses that skew those results. Which is why having a 4x larger L2 cache will be massive to productivity/performance of tasks that benefit from it. Yeah; size does matter. But not at a hit to latency. Processors run too fast to starve them for so many clock cycles...

Agreed that there is some tricks we're not privy to.

In your stated workflow/workload - I wouldn't be so worried about a weekly process that takes less than half the night to run.

If you're upgrading your system; sure - always buy as much compute as you can (like I always state...) - but to buy a system to replace what you have for a few hours weekly routine is not what I would consider a smart buy.

 

When that times comes I see if my board can still take the new AMD CPU if not I wait until it becomes affordable.

 

"Why Your Customers Are Buying Ryzen"
http://partner.amd.com/en-us/Pages/article-why-customers-are-buying-ryzen.aspx

"Ryzen™ has been an incredibly successful launch for AMD, and consumers are delighted with the performance offered by Ryzen processors. With many CPU technology advancements, Ryzen offers unparalleled performance in a multitude of tasks.

Ryzen™ 5 and Ryzen™ 7 processors target different consumer markets, but both offer incredible power when gaming, creating multimedia content, and workstation tasks (applications such as photo editing, CAD programs, etc.). As customers demand more from their systems, they search for a processor that offers them the freedom to perform various tasks, often at the same time.

Ryzen: The Next Level of Processing
..."

Ryzen 7 1700X Balance of Power Chart

 

Threadripper’s fake score-VideoCardz.com

"After playing with this for 10 minutes I realized that not only did he edit the file in cb_ranking directory, he also changed the color of the bar manually — the color is simply wrong."



The competitor Intel integrates Thunderbolt 3 in CPUs and releases specifications- Guru3d.com

"Intel will be giving the adoption rate of Thunderbolt 3 technology a boost by integrating it into upcoming 2018 processors. PC and motherboard manufacturers as such will not need to add and purchase additional components.

Royalty free or not, it however will likely drive upwards CPU prices. Thunderbolt has never seen a great deal of market penetration, but that may be about to change with Intel natively integrating Thunderbolt 3 into future CPU models and making the protocol royalty-free to chipmakers in 2018. Thunderbolt 3 comes offers transfer speeds of 40Gbps, which roughly works out to a 4K movie in 30 seconds. It can power devices, and connect to 4K peripherals. But Thunderbolt’s had six years to go mainstream, a combination of high cost and low availability have hampered its success."

 

Watch the video with the Ryzen people. When asked on the inclusion of thunderbolt on the board, they said there is no technical reason it would not work, but is left to the MB manufacturers to include on Threadripper and Epyc.

Sent from my SM-G900P using Tapatalk

 

Intel promised TB3 integration in Skylake and Kabylake too, never arrived, I'd not count on it until Intel actually ships and CPU integrated TB3 is tested in **new** motherboards that support it.

 

Doesn't matter who promised it: don't count on anything that you depend/will depend on.

The only thing that counts is the bird in the hand and not the two in the bush...

 

With all due respect I do worry about my workflow as it is done with me present and having to do the work or wait and do other things while being worked on, not an automated one run at night. And yes I am referring to cache misses being much more costly than latency.

 

For the records... From the competitor
Specs on Core i9 7900X (Skylake-X) - It'll get a High Turbo Boost Clock-Guru3d.com

Damn sorry we can't have it in laptops

Today's good one...

 

http://m.hexus.net/tech/news/cpu/106144-intel-core-i9-7900x-skylake-x-processor-spotted-database/
Guess what? SiSoft says 175W TDP, 4GHz base, 4.5 boost.

So those saying Intel is better at energy compared to the 16C Ryzen may be laughed at, if true!

Edit: same news, different source. But prior leaks have 4.5 as boost 3.0, 4.3 as boost 2.0. In other words, 4.5 on a single core took pushing the tdp WAY up. People won't OC these like the others, as it seems less headroom is likely left compared to usual, or I suspect. It's fine as those that don't OC can get more performance without the work and those numbers are still good. But you are a ninja!

Sent from my SM-G900P using Tapatalk

 

Higher clocks seems Intel's way to compete. Maybe they hope for gamers.

 

I'll take the extra threads. I am considering just getting the cheaper 12C AMD, considering the upgrade in a year. I'll be able to squeeze $400-500 out of my maximus VIII extreme and skylake ($600 is a stretch at this point), meaning between the MB and CPU, that may be the best price point if I adopt a month after release (or two months). But, Intel is not going to like what is coming for them!

I'm starting to price in a home 10Gbit Fiber network, including switch, all connectors and adapters to Ethernet termination for some devices, where needed.

Also, notice how all these rumors put Intel's 6 & 8 core chips at 28PCIe lanes. I'm betting cost cutting to try to lower price to compete with AMD's mainstream line. That means only Intel's 10 & 12C chips will have 44 lanes, which AMD's entire HEDT lineup will have on all chips. Seriously, Intel's weaknesses are being exposed!

Sent from my SM-G900P using Tapatalk

 

If you need to be present; I'm in 100% agreement then too.

~156 hours saved per year is easily justifiable towards a new system. If your time is worth ~$13/hour or more; a new ~$2K platform that can give you those time savings is easily justified.

Cache misses are more costly than mere cache latency, yeah. But even at the ns level; if every cache hit is 2x to 4x greater than the competition, then a cache size of 8x or 16x is effectively needed over what the other side is using to make things even 'theoretically' balanced.

 

Since we're just speculating; 175W TDP on it's own doesn't mean anything. It depends on the performance that will be offered (per Watt). Again; time will tell (no pun intended).

With a given workload; the time to completion (and idle) will dictate total power consumption. Knowing just the TDP doesn't indicate anything.

And most important; the actual workload involved will be of high importance too.

We just may see that AMD and Intel are matched to different workflows better (or not).

What I know for a fact? Anytime I have 'upgraded' a platform to a higher performing one (measured in multiple threads) with a decrease in the single core performance (aka 'snappiness'), it didn't feel like much of an upgrade. Even if at the end of the day I was more productive.

The point being? Both single core and multicore performance needs to be at a pinnacle for me to currently spend additional $$$$$ on new platforms.

 

As far s snappy, you may have to most of the time give this to Intel. That snappy feeling usually relates to single core speed and efficiency. Not always though but most of the time it does. Also I haven't seen a Ryzen 1800x yet but going by Cinebench single thread score 161 is better than where I am now with my mobile system I think it will be snappy enough.

 

Snappiness can be gimmicky. It affects our psychology about the system. But, if the system runs faster with more threads, the feel is nothing more than self-satisfaction. Also, software, itself, can affect perceptions of snappiness!

We are quickly approaching a point where the snappiness between generations doesn't seem to greatly increase. Because our perceptions tell us it isn't that great of a difference, then it is a non-issue. This usually is in relation to the OS and browsers that people care, as properly optimized systems for multi-threading can reduce load times for some programs.

This is why I saw practically no difference between the 6700K and the 4790K on snappiness. Because of that, it will be about the same between Intel's new CPU and Ryzen, which has the IPC of BW-E and not that different from Intel's current offerings.

Sent from my SM-G900P using Tapatalk

 

Snappiness is not gimmicky to me. With a snappy enough system; I am not concentrating on the mechanics of my input - rather, I'm just flowing with the ideas as they come...

While we can pretty accurately judge how snappy two processors from the same manufacturer can (or at least 'should') be, the same cannot be said between processors from different manufacturers. Again; the whole platform is at play there and simple, one number 'scores' like IPC, clock speed, etc. cannot convey what the whole system is capable of producing - for the user.

Before SSD's, where notebooks were concerned; there was either a snappy system (short stroked drive - 5-10% of total capacity) or there was a system where you could do real work on.

Today with an SSD powered system, a QC or better CPU and 16GB/32GB/64GB of RAM or more, we have both snappiness and a platform where real work happens too. At this point, I wouldn't trade either for love or money.

 

I just somehow doubt with a good SSD and a Threadripper with quad channel memory that I am going to want for more snap. Just a feeling mind you.

 

No doubt!

Just thinking forward a couple of years into DIMM powered XPoint 'storage/ram'. If/when it comes it will definitely reset my expectations for what 'snappy' really means to me.

Of course, in a couple of years AMD may have access to a flavour of XPoint from other manufacturers too.

 

Yeah, I can't buy for 2-4 years down the line. As it is today I do not need a system. Now that I have the 3440x1440 display my work flow is just awesome, those video encodes etc. can just sit there and do their thing. With the quad core as it is I can give up some speed and watch video's/movies browse the web and work seamlessly.

So I am pretty sure I am good for the next year or two. My reason for looking now is to get in while Windows 7 is still an option and start gearing for a Linux future. Before all this Ryzen stuff I was just going to use what I have eventually under Linux.

 

My desktop is only a year old, but I'll take the hit to get on Ryzen (that way it is either adding ram, changing CPU with the 7nm chip, if significant performance boost, and updating graphics as I go). While working, I can get it where a quad core isn't snappy, and it isn't just memory resources (I need around 20+ GB of ram to do my work, but have 32GB; it's more all the programs big down the snappy feel, not to say it is horrible).

This also feeds into my long-term goals of a home server, so I'll feed my dreams and hobbies.

Finally, AMD is working hard for working well with Linux. If you check Geekbench, both Intel and AMD perform better in it under Linux than Windows. But, I get it. I just hope the TR drivers supplied by AMD contain what is needed for better Win7 support by them (like they had to at first with Ryzen 7).

Sent from my SM-G900P using Tapatalk

Edit: it just clicked! We are literally less than a week away from computex goodness and so much information!!! If they show it off day one, we have 5 days. If they wait for the end, a bit longer, but not much. Good thing my RMA replacement keyboard gets here Tuesday! ;-) might be typing furiously all over the interwebs! (Only moderated by my nieces and nephew visiting, which are young enough to need some attention)...

 

@ajc9988

Look's like some MOBO makers are paying a little more attention to Ryzen. MSI is going to launch another high end X370 board and that's before Threadripper. Perhaps there is some money to be made with Ryzen.

http://www.tomshardware.com/news/msi-x370-gaming-m7-motherboard,34520.html

EDIT: I would have created a Ryzen 7 build if it weren't for Threadripper and I still may to see how it compares to my X99 builds. Threadripper will probably be in my next home workstation.

 

That is nice to see! I'm betting added bclk oc and may have qualified it for the AGESA 1.0.0.6 to help with the ram speed support. But that is refreshing to see.

But, once I knew of Threadripper, or rather suspected a larger number of cores and quad channel, I had to wait. Otherwise Ryzen would already be in my desktop. Lol.

But, if I build another mainstream build anytime soon, I'd likely go Ryzen. For laptop, Clevo still has my heart and a coffee lake laptop with 6 cores would be where my eye is on that front. If MSI created a desktop chip laptop with a 17" display, then that is where I'd go. But, that is because mobile cpus don't do it for me and I haven't heard of a 1600X laptop planned (AM4 socket, but 6C might be right on heat, etc., although an 8C...).

Sounds like we are all anxious! Let's hope Intel and AMD use the same benches at Computex (or allow journalists to run whatever they want on the sample machines). AMD may do an LN2 demo like last time, also, and take some golds. Intel will take some on core count bases as well. It's going to be exciting!!!

Sent from my SM-G900P using Tapatalk

 

AMD Talks AGESA v1.0.0.6 and DRAM-Guru3d.com

 

Community Update #4: Let's Talk DRAM!

Posted by rhallock in Gaming on May 25, 2017 2:57:54 PM

"Welcome to the fourth installment of the Ryzen Community Updates! If you’re checking into this series for the first time, this is where we let our community know about all the exciting updates that are on their way for the AMD Ryzen™ processor.

We’ve covered a lot of ground in a short amount of time: game updates, new Windows® power plans, stability improvements, feature additions and much more. Today’s update is the one many of you have been most excited about: the AGESA that focuses on overclocked memory. There’s some great news for IOMMU/ACS users, too!"

Spoiler: Details...

What is an “AGESA”?

AGESA is an acronym that stands for “AMD Generic Encapsulated System Architecture.” As a brief primer, the AGESA is responsible for initializing AMD x86-64 processors during boot time, acting as something of a “nucleus” for the BIOS for your motherboard. Motherboard vendors take the core capabilities of our AGESA updates and build on them with their own “secret sauce” to create the BIOS files you download and flash. Today, the BIOS files for AMD AM4 motherboards are largely based on AGESA version 1.0.0.4.

Beginning this month, as we promised to you, we began beta testing a new AGESA (v1.0.0.6) that is largely focused on aiding the stability of overclocked DRAM (>DDR4-2667). We are now at the point where that testing can begin transitioning into release candidate and/or production BIOSes for you to download. Depending on the QA/testing practices of your motherboard vendor, full BIOSes based on this code could be available for your motherboard starting in mid to late June. Some customers may already be in luck, however, as there are motherboards—like my Gigabyte GA-AX370-Gaming5 and ASUS Crosshair VI—that already have public betas.

Starting With Virtualization
If you’re the kind of user that just needs (or loves!) virtualization every day, then AGESA 1.0.0.6-based firmware will be a blessing for you thanks to fresh support for PCI Express® Access Control Services (ACS). ACS primarily enables support for manual assignment of PCIe® graphics cards within logical containers called “IOMMU groups.” The hardware resources of an IOMMU group can then be dedicated to a virtual machine.

This capability is especially useful for users that want 3D-accelerated graphics inside a virtual machine. With ACS support, it is possible to split a 2-GPU system such that a host Linux® OS and a Windows VM both have a dedicated graphics cards. The virtual machine can access all the capabilities of the dedicated GPU, and run games inside the virtual machine at near-native performance.

This is certainly a complicated setup for most users, but I have no doubt that there will be a whole lot of you enthusiastically nodding at this news. We’re grateful for your feedback and your patience, and we hope the new support for ACS serves you well.

What's Next For Memory
AGESA 1.0.0.6 officially adds 26 new parameters that can improve the compatibility and reliability of DRAM, especially for memory that does not follow the industry-standard JEDEC specifications (e.g. faster than 2667, manual overclocking, or XMP2 profiles).

The following table spells out all the new parameters, and provides a few words on what they do. Keeping in mind that this is overclocking territory, manual or automated control of these parameters should nevertheless make it a little more straightforward to use DDR4-3200 modules—or faster if you have the talent!

Spoiler: text...

Parameter Function Values
Memory clocks Added dividers for memory clocks up to DDR4-4000 without refclk adjustment. Please note that values greater than DDR4-2667 is overclocking. Your mileage may vary (as noted by our big overclocking wartning at the end of this blog). 133.33MT/s intervals (2667, 2933, 3067, 3200, 3333, 3466, 3600, 3733, 3866, 4000)
Command rate (CR) The amount of time, in cycles, between when a DRAM chip is selected and a command is executed. 2T CR can be very beneficial for stability with high memory clocks, or for 4-DIMM configurations. 2T, 1T
ProcODT (CPU on-die termination) A resistance value, in ohms, that determines how a completed memory signal is terminated. Higher values can help stabilize higher data rates. Values in the range of 60-96 can prove helpful. Integer values (ohms)
tWCL/tWL/tCWL CAS Write Latency, or the amount of time it takes to write to the open memory bank. WCL is generally configured equal to CAS or CAS-1. This can be a significant timing for stability, and lower values often prove better. Integer values (cycles)
tRC Row cycle time, or the number of clock cycles required for a memory row to complete a full operational cycle. Lower values can notably improve performance, but should not be set lower than tRP+tRAS for stability reasons. Integer values (cycles)
tFAW Four activation window, or the time that must elapse before new memory banks can be activated after four ACTIVATE commands have been issued. Configured to a minumum 4x tRRD_S, but values >8x tRRD_S are often used for stability. Integer values (ns)
tWR Write recovery time, or the time that must elapse between a valid write operation and the precharging of another bank. Higher values are often beneficial for stability, and values < 8 can quickly corrupt data stored in RAM. Integer values (ns)
CLDO_VDDP
Voltage for the DDR4 PHY on the SoC. Somewhat counterintuitively, lowering VDDP can often be more beneficial for stability than raising CLDO_VDDP. Advanced overclockers should also know that altering CLDO_VDDP can move or resolve memory holes. Small changes to VDDP can have a big effect, and VDDP should not be set to a value greater than VDIMM-0.1V. A cold reboot is required if you alter this voltage.

Sidenote: pre-1.0.0.6 BIOSes may also have an entry labeled “VDDP” that alters the external voltage level sent to the CPU VDDP pins. This is not the same parameter as CLDO_VDDP in AGESA 1.0.0.6.

Integer values (V)
tRDWR / tWRRD Read-to-write and write-to-read latency, or the time that must elapse between issuing sequential read/write or write/read commands. Integer values (cycles)
tRDRD / tWRWR Read-to-read and write-to-write latency, or the time between sequential read or write requests (e.g. DIMM-to-DIMM, or across ranks). Lower values can significantly improve DRAM throughput, but high memory clocks often demand relaxed timings. Integer values (cycles)
Geardown Mode Allows the DRAM device to run off its internally-generated ½ rate clock for latching on the command or address buses. ON is the default for speeds greater than DDR4-2667, however the benefit of ON vs. OFF will vary from memory kit to memory kit. Enabling Geardown Mode will override your current command rate. On/Off
Rtt Controls the performance of DRAM internal termination resistors during nominal, write, and park states. Nom(inal), WR(ite), and Park integers (ohms)
tMAW Maximum activation window, or the maximum number of times a DRAM row can be activated before adjacent memory rows must be refreshed to preserve data. Integer values (cycles)
tMAC Maximum activate count, or the number of times a row is activated by the system before adjacent row refresh. Must be equal to or less than tMAW. Integer values (cycles)
tRFC Refresh cycle time, or the time it takes for the memory to read and re-write information to the same DRAM cell for the purposes of preserving information. This is typically a timing automatically derived from other values. Integer values (cycles)
tRFC2 Refresh cycle time for double frequency (2x) mode. This is typically a timing automatically derived from other values. Integer values (cycles)
tRFC4 Refresh cycle time for quad frequency (4x) mode. This is typically a timing automatically derived from other values. Integer values (cycles)
tRRD_S Activate to activate delay (short), or the number of clock cycles between activate commands in a different bank group. Integer values (cycles)
tRRD_L Activate to activate delay (long), or the number of clock cycles between activate commands in the same bank group. Integer values (cycles)
tWR Write recovery time, or the time that must elapse between a valid write operation and the precharging of another bank. Higher values are often better for stability. Integer values (ns)
tWTR_S Write to read delay (short), or the time between a write transaction and read command on a different bank group. Integer values (cycles)
tWTR_L Write to read delay (long), or the time between a write transaction and read command on the same bank group. Integer values (cycles)
tRTP Read to precharge time, or the number of clock cycles between a READ command to a row and a precharge command to the same rank. Integer values (cycles)
DRAM Power Down Can modestly save system power, at the expense of higher DRAM latency, by putting DRAM into a quiescent state after a period of inactivity. On/Off
Until next time
What are you interested in hearing more about in our next AMD Ryzen Community Update? Let us know on Twitter @AMDRyzen!

Robert Hallock is a technical marketing guy for AMD's CPU division. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.
1. WARNING: Overclocking memory will void any applicable AMD product warranty, even if such overclocking is enabled via AMD hardware and/or software. This may also void warranties offered by the system manufacturer or retailer or motherboard vendor. Users assume all risks and liabilities that may arise out of overclocking memory, including, without limitation, failure of or damage to RAM/hardware, reduced system performance and/or data loss, corruption or vulnerability. GD-112

 

https://www.pcgamesn.com/amd/amd-kyzen


https://newsworld.co/8gb-memory-hbm2-for-a-single-card-radeon-rx-vega-complete-amd-at-160/
http://www.guru3d.com/news-story/amd-rx-vega-hbm-2-8gb-memory-stack-costs-160,2.html

 

I know AMD bet big on HBM technologies but $160 per 8GB (x2 stacks) is a hell of lot to pay at the manufacturer level. I know HBM2 is going on the high end cards but I can't see a card that can stand toe to toe with a 1080 TI costing less than the 1080 TI with this memory. If this is the case, the 1080 TI will continue to win marketshare in the top end.

 

16gb is on the titan competitor. 8GB is on the Ti competitor. So, it is $160 for the ram. That means, to reach $600 with 30% margin, you would need $600-180-160=$260 for the cost of the rest of the card.

Sent from my SM-G900P using Tapatalk

 

If the reference card is $600 and the aftermarket cooler variances add another $50 to $75, you're near the 1080 TI level. I'm sure during the summer, there will be rebates for 1080 TI cards, pushing those lower.

The problem I foresee is Volta is right around the corner. Nvidia is once again going with brute strength and while the Vega series like most AMD cards should age very well, they won't be able to compete with Volta.

This speculation means nothing if AMD has something waiting in the wings for Volta, but it seems like we are going to have a Big Brother, Little Brother situation all over again. I really want to build an all AMD rig but AMD put themselves in a weird spot. Essentially, they are competing against two generations of Nvidia GPU's.

 

First, it isn't brute force. The HBCC causes the minimum frame rate to be much higher. Unless Nvidia has got something equivalent, that is one for Vega.

Second, Nvidia Volta this year would use either 12 or 14gbps, not the 16 which isn't available until next year. This still gives an effective refresh higher than AMD's if the 14 is used, but doesn't guarantee better minimums.

Now, next year is 14nm+ Vega 20 going against the full Volta lineup. Volta is supposedly 12nm. Then Navi goes 7nm and Nvidia will likely do the same.

But don't count AMD out until you get the full numbers.

Sent from my SM-G900P using Tapatalk

 

Just to many, what ifs, should haves, will be's, kind of's, by then's I am so confused.

 

Once they said that Vega for consumers won't be here for months, I tuned out. If Nvidia shows up in August with Volta XX80 and XX70, then it is best to wait and just see them head to head.

 

Just kidding. , I follow it but just saying there are so many variables it is almost too far out there. Exciting times to see the standards being challenged but it may be up to the end of the year before enough plays out to see where the markets lie.

As an example the below link suggest possibly August

http://www.isportstimes.com/article...rds-launching-soon-after-frontier-edition.htm

 

But, my thought was, since it didn't drop in May and won't in June, I'll stick with my 980 Ti and focus on the MB and CPU upgrade this summer.

Let's get real, everyone in hear watches like a hawk! LOL!

 

http://www.techradar.com/news/computex-2017-what-to-expect-from-the-years-biggest-computing-show
https://www.techporn.ph/amd-raven-ridge-apu-ryzen-vega-will-come-to-notebooks-soon/

 

Full Intel X299 HEDT Skylake X, Kaby Lake X and Coffee Lake S Mainstream Platform Details Unveiled – Core i7-7920X Skylake X Flagship, Core i7 7740X Kaby Lake X Flagship, Coffee Lake To Get 6 Core and 4 Core CPUs-Wccftech.com
"Aside from that, Intel X299 has full support for CPU overclocking which is a bonus, we just hope that the CPUs are running cooler compared to the Kaby Lake S parts where Intel had to ask users to refrain from overclocking to avoid overheating issues."

Intel Coffee Lake S CPU Lineup To Feature Hyperthreaded Core i7 and Core i5 Models With Up To 6 Cores, Core i3 To Boast 4 Cores – Support Recommended on Z370 Platform With LGA 1151 V2 Socket
"The conference also detailed a bunch of new information on the upcoming Coffee Lake CPUs. The Coffee Lake S family which is the codename for the mainstream desktop platform is already known to feature the first 6 core mainstream processor from Intel. The processor lineup will launch in the third quarter with a release focused around August at Gamescom 2017. The release will include a series of 6 and 4 core models."

Coffee Lake S will be launching alongside the brand new Z370 chipset which is known as CFL-PCH (Coffee Lake PCH). The platform will feature the LGA 1151 V2 socket so while it will have the same number of pins as the current LGA 1151 socket, it seems like Intel is moving to ditch support of the Coffee Lake CPUs on older LGA 1151 socketed motherboards such as Z270 series. It is stated that Intel will clearly be differentiating the socket with the use of “V2” moniker so consumers don’t end up buying the wrong board for their Coffee Lake desktop processor." And the new Intel Core i5 will be the old i7

 

This demonstrates even more how they were milking everyone... Now they just pull the switch and increase everything (cores, HT) trying to push AMD again out of the market to start milking again.

 

Intel Who?

!!Stop buying Intel!!

Let's all give Intel a good long 5 year "timeout"

 

Let's go and get a Threadripper!
I really have to rethink the idea of having one monstrous server. Then, streaming everything from it to the clients (Tablets, small notebooks). Does this work good now with Win10? I remember seeing Linus trying such a build but it was still a bit "itchy".

Either way, this 12c Threadripper looks very promising!

edit:
http://wccftech.com/amd-threadripper-1998x-and-threadripper-1998-processors-x399-x390/

*Want to have intensifies*

 

I agree. Also, remember this uses Zeppelin cores, which have a faster 100Gbps data fabric between cores and for memory than Ryzen CCX, meaning even Ryzen may not be a full gauge for performance.

I agree!

Pretty much meet my expectations. But, if the Q3 coffee release is true, they just fully cannibalized the Kaby-X chips. The 14nm++ is supposed to be the same IPC gain as between skylake and Kaby. Plus, with 6 cores...

Sent from my SM-G900P using Tapatalk

 

That is Intel saying we are not done just yet. It seems they are trying to soften the inevitable blow Threadripper and Epyc will soon have. While on gaming they will still be king the lucrative high end of the HEDT, work station and server markets are about to take a sever hit.

 

More from the competitor...
Computex 2017: Intel To Release 18 core Core i9-7980XE Skylake-X CPU-Guru3d.com
Yeah, Amd has pushed some changes

"I stated it before, there's a bit of a panic reaction from Intel towards Ryzen 16-core parts, hence the 18-core models seem just that, a reaction to offer something more opposed to AMD Threadripper. These are HUGE steps for Intel, something we have not seen in years."

 

so, let's bet on Xeon type pricing. Also, since these are basically Xeon chips to get the higher core count, it is a matter of them just tweaking it to get a higher speed, then unlocking the chip for overclocking. The real question remains whether they can get the speeds with those cores.

This looks entirely reactionary and if not ready for August, when the 12-core is planned, then they leave AMD to mop up. If the pricing looks like a Xeon, then AMD will mop up. If AMD makes an Epyc 16 core that is overclockable, then AMD will mop up (this will beat the 18-core only due to the extra PCIe lanes and potentially faster transfer speeds if the entire board incorporates the 100Gbps data fabric). Considering mid-4GHz is expected with the 12-core, with 16-cores, you could see 4GHz, thereby being only an IPC advantage. If the speed is less, like the Xeon series, then the IPC advantage may not be enough.

I read a comment on that post that Intel isn't scared. This says otherwise. Unless pricing is right, it really means nothing.

Big round of applause for AMD for kicking Intel in the nuts!

Sent from my SM-G900P using Tapatalk

 

These are likely to be more related to the slow clock rate Xeon chips if so at such high core counts. This again goes back to the old days when they were competing with one another and vaporware was everywhere. AMD has not even put out the benchmarks yet and Intel is in panic mode, something great may be coming forth in short order.

 

http://hexus.net/tech/news/cpu/1062...s-include-18-core-flagship-says-leaked-slide/
According to Hexus, the keynote from Intel is scheduled for tomorrow at 2PM taiwan. So, we will find out soon.

The reason I say higher clock speed is the X series chipset is the same, but they always clock it higher than the same core count equivalent Xeon. But, there are limits, and as Intel's core count increases, the speed drops a lot. So, the overclocking on it, if it comes to the same, is why IPC will likely be the difference between the two.

Now, previous leaks put the release of the higher core count Xeons this summer, by August. This is why, if that holds true, they are doing a separate run on these other chips, which suggest slight tweaks to try to reach higher speeds.

To be fair, the rumors surrounding Epyc suggest lower speeds for its Opterons as well, compared to Ryzen and Threadripper. So, we will find out soon.

AMD's presentation will be shortly after:
https://www.amd.com/en/events/computex-2017

 

Even if they announce high clock rates it still has to go by the I'll believe it when I actually see it meaning real world silicon. Intel can white paper and announce all they want, it is hard to argue with hardware you can hold in your hand and play with if you have the bank.

Edit; Wow that is a late conference on the 30th.

 

I just looked it up. Taiwan is 12 hours ahead of EST. So Intel's presentation is at 2 on Tuesday Taiwan time (2AM tonight)! That means the AMD presentation is the day after in Taiwan in the morning. I haven't looked up Nvidia's, but they said it would have info on Volta today, meaning they may present before Intel even. I'll check in a little bit. Looks like today and tomorrow will have some large headlines.

Then comes news of motherboards. Intel will have HEDT on an mITX board. Makes me wonder if AMD would have the same. We could be looking at smaller workstations available (or that can be built if cooling is sufficient).
http://www.tweaktown.com/news/57771...x.html?utm_source=dlvr.it&utm_medium=facebook

Sent from my SM-G900P using Tapatalk

 

Hey guys, I'm trying to collect all the Computex 2017 stuff here:

Computex 2017 - Events, Announcements, Video, Articles...

Please post what you find when you find it, I'll do my best to keep up, but I'm not able to do it 24/7

It will be nice to have a thread to find all the announcements from the show as a reference moving forward.

Please continue to post AMD content here too...