Weird RAM Timings

[sphinx64]

I have 4 sticks of Corsair Vengeance LED CMU16GX4M2C3200C16 with a total of 32Gb RAM
RAM is now runnıng at 2880MHz (confirmed with AIDA, CPUz, Bios, etc.) with Bios settings (9945);
DOCP set to 4
BCLK Frequency=120.000
Memory rating set to 2880MHz (actually there was this option instead of 2933)
Core Performance Boost=enabled
CPU Core Voltage=auto
CPU SOC Voltage=manual
- VDDSOC Voltage Override=1.1
DRAM Voltage=1.35
DRAM Boot Voltage=1.35
DRAM CL=16
DRAM RCD=18
DRAM RP=18
DRAM RAS=36
DRAM RC-RFC1-RFC2-RFC4-RRDL-RRDS-FAW=all auto (manual entry causes instability - all of the values obtained by AIDA and Thaiphoon Burner)
ProcODT_SM=68.6
Cmd2T=auto (either 1T or 2T causes instability)
Gear Down Mode=auto
Power Down Enable=auto
CLDO_VDDP=auto (no matter how hard I tried to follow Elmor's instructions properly, it wouldn't stick)
Rest of the Bios values are at default.
Memory passes all the tests and sytem is stable, however I am a bit nervous of the 2880MHz, as I never heard of such a weird #. Any ideas or suggestions?

[entropic-remnan]

Praz wrote:
Hello

The preset memory divider speeds are based on a 100MHz BCLK speed. Changing the BCLK speed will also alter the memory speed for a given divider. That 2933MHz is no longer an option with the BCLK speed set to 120Mhz should be expected. Memory is spec'd to run as sold. When combining kits one should plan on manual tuning and/or reduced speed or timings. The Spread Spectrum setting is for the switching frequency of the VRM circuit. Its chosen setting will have no influence on BCLK frequency but for other reasons is best set to disabled when overclocking.

That makes sense. Some folks said that they could see the spread spectrum modulation in the Bclk in monitoring tools but it always jitters in monitoring tools so I figured that might not be true. Thanks for clarifying.

[sphinx64]

Thanks to Praz and entropic-remnants. I decided that weird RAM setting is not a good way to go and now back to 2666. With 4 sticks of non-compatible RAM this is what I can get at most. Funny thing is that this memory is listed in the QVL ASUS listed but it simply is not compatible. M2N-E, M2N-E SLI, Crosshairs III-IV-V never had these issues; everything worked out of the box with overclocking of the CPU and memory without a hassle. Therefore, this is not ASUS but AMD to blame. They serviced a non-ready platform and we paid good amount of money to this system, which manufacturers had no tools in their hands to make their products work. We are still months away from a real solution where we can have our QVL listed memory work in stock speeds and timings.
I can only criticize ASUS for only one thing; back in old days there were these official ASUS forums, where people go and post their problems and get their solutions. When I log into official ASUS forum (by the way it is, as probably most of you don't know; http://vip.asus.com/forum/topic.aspx?board_id=1&model=CROSSHAIR+VI+HERO&SLanguage=en-us) there are 9 TOTAL posts with at the most 1000 views in total; a ridiculous # for e new board with lots of problems. May be times have changed, but it is sad to see the condition of official forums. If people began to get their information from multiple sources, there will always be data corruption, resulting in customer dissatisfaction.

Anyways,

Thanks Praz and entropic-remnants for your incredible dedication on helping.

Regards

[Praz]

sphinx64 wrote:
Thanks to Praz and entropic-remnants. I decided that weird RAM setting is not a good way to go and now back to 2666. With 4 sticks of non-compatible RAM this is what I can get at most. Funny thing is that this memory is listed in the QVL ASUS listed but it simply is not compatible. M2N-E, M2N-E SLI, Crosshairs III-IV-V never had these issues; everything worked out of the box with overclocking of the CPU and memory without a hassle. Therefore, this is not ASUS but AMD to blame. They serviced a non-ready platform and we paid good amount of money to this system, which manufacturers had no tools in their hands to make their products work. We are still months away from a real solution where we can have our QVL listed memory work in stock speeds and timings.
I can only criticize ASUS for only one thing; back in old days there were these official ASUS forums, where people go and post their problems and get their solutions. When I log into official ASUS forum (by the way it is, as probably most of you don't know; http://vip.asus.com/forum/topic.aspx?board_id=1&model=CROSSHAIR+VI+HERO&SLanguage=en-us) there are 9 TOTAL posts with at the most 1000 views in total; a ridiculous # for e new board with lots of problems. May be times have changed, but it is sad to see the condition of official forums. If people began to get their information from multiple sources, there will always be data corruption, resulting in customer dissatisfaction.

Anyways,

Thanks Praz and entropic-remnants for your incredible dedication on helping.

Regards

Hello

You're welcome. The VIP forum you linked to is strictly a user-to-user forum with no official support from ASUS. It has always been this way and there should be a statement confirming this somewhere on that forum.

It seems AMD took a gamble and it will most likely pay off. They produced a product at half the price point of Intel which could also outperform Intel in CINEBENCH. But the rest seems as you point out as being rushed to market and in my opinion overall sub-par when compared to Intel. In AMD's favor the majority of these processors sold will never be overclocked so the shortcomings will not be known by the user and the overall cost of ownership will appear to be a bargain. There's an old saying that you get what you pay for and Ryzen compared to Intel is the perfect example. Hopefully Intel does not move in the same direction in order to compete at the same pricing.

I have screenshots posted of full stability with the memory at 3600MHz and general usage at 3733MHz. However, recognizing what my particular CPU is capable of my daily settings are 3200MHz memory speed. A person needs to be aware of what constraints are imposed by the individual CPU and work within those confines. After all one can throw a turd at the wall to see if it will stick only so many times.

[danjw]

Praz wrote:

It seems AMD took a gamble and it will most likely pay off. They produced a product at half the price point of Intel which could also outperform Intel in CINEBENCH. But the rest seems as you point out as being rushed to market and in my opinion overall sub-par when compared to Intel. In AMD's favor the majority of these processors sold will never be overclocked so the shortcomings will not be known by the user and the overall cost of ownership will appear to be a bargain. There's an old saying that you get what you pay for and Ryzen compared to Intel is the perfect example. Hopefully Intel does not move in the same direction in order to compete at the same pricing.

This isn't entirely fair to AMD, Intel had similar problems with the launch of their Broadwell-X platform. Any company can have executives decide to ship a product that isn't quite ready. That may be coming again with Intel pushing product launches forward to respond to AMD.

[entropic-remnan]

How is it not "fair" to AMD, lol? Those executives are PART of AMD, and it exists only with them whether we like them or not.

If you mean it's not fair to the engineering department of AMD, fair enough. But the company as a unit takes the brunt of the criticism and rightly so. Anyone who handed over their money for a processor with AMD on it, and a motherboard that resulted from their decision making process, has every right to criticize.

Don't let them off the hook. They need to execute and give us what we paid for and trusted them for. I'm not buying any "fan" stuff here.

Oh, yeah, I've had Intel processors a LONG time so we can talk about Intel's foolishness, lol. But Intel has gotten DDR4 down to a science and AMD/Asus aren't even doing the BASICS of reading the XMP properly. That's not reassuring. How many BIOS have there been now?

That's my take. I don't want my money back but I do want to see this made right soon and at least the basics nailed down.

[sphinx64]

entropic-remnants wrote:
How is it not "fair" to AMD, lol? Those executives are PART of AMD, and it exists only with them whether we like them or not.

If you mean it's not fair to the engineering department of AMD, fair enough. But the company as a unit takes the brunt of the criticism and rightly so. Anyone who handed over their money for a processor with AMD on it, and a motherboard that resulted from their decision making process, has every right to criticize.

Don't let them off the hook. They need to execute and give us what we paid for and trusted them for. I'm not buying any "fan" stuff here.

Oh, yeah, I've had Intel processors a LONG time so we can talk about Intel's foolishness, lol. But Intel has gotten DDR4 down to a science and AMD/Asus aren't even doing the BASICS of reading the XMP properly. That's not reassuring. How many BIOS have there been now?

That's my take. I don't want my money back but I do want to see this made right soon and at least the basics nailed down.

+1 to the righteous, to the point comment. I always used AMD platforms since 486SX days and this company was and used to be a trusted company when it came down to launching new platforms or products. Kudos to their development, pricing, and value they offer; however, we are not lab rats testing their non-ready platforms. It is not an excuse saying that Intel had similar problems during product launches because IMO, they only value profits and market share, but not the end user. AMD was not used to be this type of company in the past, and hope that they will not turn into one.

[danjw]

entropic-remnants wrote:
How is it not "fair" to AMD, lol? Those executives are PART of AMD, and it exists only with them whether we like them or not.

If you mean it's not fair to the engineering department of AMD, fair enough. But the company as a unit takes the brunt of the criticism and rightly so. Anyone who handed over their money for a processor with AMD on it, and a motherboard that resulted from their decision making process, has every right to criticize.

Don't let them off the hook. They need to execute and give us what we paid for and trusted them for. I'm not buying any "fan" stuff here.

Oh, yeah, I've had Intel processors a LONG time so we can talk about Intel's foolishness, lol. But Intel has gotten DDR4 down to a science and AMD/Asus aren't even doing the BASICS of reading the XMP properly. That's not reassuring. How many BIOS have there been now?

That's my take. I don't want my money back but I do want to see this made right soon and at least the basics nailed down.

I wasn't saying AMD was right, I was saying Intel has been just as bad. 😉

[entropic-remnan]

danjw wrote:
I wasn't saying AMD was right, I was saying Intel has been just as bad. 😉

Ha ha! Well, sorry then, brother -- there's so much fan-boy crap that goes on in these forums and I have little tolerance. I mistook you for going in that direction.