have you set to DHCP ? or static IP ? is the subnet mask correct ? your PC must also have IP like 192.168.1.x
BMC switch'ed to on and done a Power cycle ?
don't think its normal. "the red light" ? can you explain !
W790 Motherboard discussion thread
- Thread starter heromode
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
BMC switch'ed to on and done a Power cycle ?
type in "https://192.168.1.13" to get in to ipmi in the browser, without the https it won't resolve. same thing happened to me.
you're battery voltage is low, mine is showing 3.136 v, this is probably the cause of the red/orange flashing led.
I'd try taking the battery out and reseating it; might be a bad contact with battery but likely it's just a drained battery.
It's not that bad, the worse thing you'll experience is loss of BIOS settings after removing power from the machine.
I do not know where on this board the battery is not visually visible. BIOS settings seem to be saved after the computer is turned off. I received the payment yesterday, the battery may be dead. In any case, thanks for the advice, you helped me a lot.
Good to hear.
I haven't updated my BMC yet but I hear the update fixes .ISO redirection problems the initial BMC release had.
Question: Assume 2400 cpu.
If attached to the Sage, it has the dual intel 10gb nics.
What if one used the Ace and added an intel x550-T2 or maybe an x710-TL2 (forget the other built-in nics)
Pricing of the 2 choices is comparable.
Obviously, there are less pcie lanes on the 2400 (and the Ace for that matter) which is no problem in this scenario as there are already plenty of lanes, but the nic would be in a pcie slot with direct cpu access. As I understand it, the Sage built-in nic would be connecting back through DMI 4. Would that make a difference compared to going directly to the cpu (albeit probably not too much) with a nic in a pcie slot?
If attached to the Sage, it has the dual intel 10gb nics.
What if one used the Ace and added an intel x550-T2 or maybe an x710-TL2 (forget the other built-in nics)
Pricing of the 2 choices is comparable.
Obviously, there are less pcie lanes on the 2400 (and the Ace for that matter) which is no problem in this scenario as there are already plenty of lanes, but the nic would be in a pcie slot with direct cpu access. As I understand it, the Sage built-in nic would be connecting back through DMI 4. Would that make a difference compared to going directly to the cpu (albeit probably not too much) with a nic in a pcie slot?
I got the ASRock W790 WS & Asus PRO WS W790-ACE, Xeon W7-2495X, and The G.Skill Zeta R5 DDR5-6400 CL32-39-39-102 1.40V 64GB (4x16GB).
The ASRock W790 WS has the following salient features over the others:
Notes on the Asus PRO WS W790-ACE:
The ASRock W790 WS has the following salient features over the others:
- 2x Thunderbolt 4 ports.
- 4x CPU PCIe M.2 slots so you don't need a carrier card.
- PCIe slot 1 is double-spaced so the GPU doesn't waste a slot. The slot is in-line with the GPU inlet / outlet ports on most distro plates.
- 4x usable CPU PCIe slots, plus an x4 PCH slot (x16/x8/x16/x4/x8), so you can use two x16 dual-slot GPUs without blocking any CPU slots. The last slot can accept a third x16 dual-slot GPU if you can fit it at the bottom of the case without impacting the connectors.
- 3x internal USB 2.0 (1 full header and 1 half header) ports for water cooling and RGB equipment.
- 4x internal USB 3.0 ports to fully populate front panel USB ports (most boards have only 2x USB 3.0 ports now).
- BIOS defaults are almost optimal and has everything turned on except CPU overclocking.
Per Core (OC ratio based on number of active cores) overclocking is buggy. If you don't use all 8 registers and zero out the unused ones, the OC settings stop working permanently. You have to CMOS reset to get it working again. A workaround is to enter fake or duplicate values and populate all 8 registers. E.g. on my 24-core CPU:
2c 48x
4c 47x
24c 46x
28c 45x
32c 44x
40c 43x
48c 42x
56c 41xACPI CPPC does not work. The BIOS reports a "Highest Performance" value of 255 for all cores regardless of the core's OC ratio. Windows cannot assign threads to the fastest cores.- Secure Boot has PH Randomization turned on by default which is incorrect. It must be turned off or Secure Boot / BitLocker may not work properly.
Windows boot sometimes BSODs with DRIVER_VERIFIER_DMA_VIOLATION. This may be related to Thunderbolt. It happens even when the ports are empty.
Notes on the Asus PRO WS W790-ACE:
- 5x total CPU PCIe slots (x16/x16/x16/x8/x8), with 4x usable CPU PCIe x16 slots when using a dual-slot GPU (x16/x16/x16/x0/x16).. The GPU must be in slot 3 to not block another x16 slot. A second dual-slot GPU will block an additional slot, unless you can fit it in the last slot at the bottom of the case without impacting the connectors.
- PCIe slot 1 is lowered and doesn't line up with GPU inlet / outlet on most distro plates. Slot 3 is even lower.
- Optional IPMI add-in card. The SAGE board has built-in IPMI.
- 8x rear USB 2.0 ports.
- 4x internal USB 2.0 ports (2 full headers) for water cooling and RGB equipment.
- Only 2x internal USB 3.0 ports. Cannot fully populate front panel USB ports.
- POST code display blocked by PCIe card.
Last edited:
"ECSM also brings up seemingly new information with an alleged Sapphire Rapids Refresh lined up for early 2024—suggesting that updated HEDT Xeon W2500 and W3500 series processors are incoming. The Intel W790 chipset should be able to run this rumored replacement lineup."
Report Suggests Intel Refreshing "Sapphire Rapids" - Updated Xeon-W Slated for Early 2024
Report Suggests Intel Refreshing "Sapphire Rapids" - Updated Xeon-W Slated for Early 2024
Level1Techs has had a few videos on this platform in recent weeks. It confirms if you want decent all core performance, some level of overclocking is needed and you need to budget 400-500W (minimum) for CPU cooling. See their Falcon Northwest system review for a "factory OC" that seems reasonable.
On the bright side, I would not have guessed a few years ago EXPO style high performance Registered ECC DIMMs would be a thing and yet here we are.
On the bright side, I would not have guessed a few years ago EXPO style high performance Registered ECC DIMMs would be a thing and yet here we are.
All CPU cores have stable factory VID programmed up to the Turbo Boost Max 3.0 ratio of the two fastest cores. You can set all cores to the maximum ratio (48x on the W7-2495X) with appropriate AVX offsets and it is 100% stable on stock voltage.
The memory always runs at command rate 1T. I cannot get the G.Skill 6400C32 4x16GB memory stable above 6000, but it can do 6000C28 100% stable. The mainstream desktop platform is able to run 6400+ only at command rate 2T, so 6000C28 at 1T is actually a very good OC. Since these are R-DIMMs, the memory controller voltage makes no difference. The registered buffer on the DIMM takes the load off the memory controller. You can run stock memory controller voltage when overclocking for lower CPU temps. The register also adds about 30ns of latency compared to mainstream desktop.
The memory always runs at command rate 1T. I cannot get the G.Skill 6400C32 4x16GB memory stable above 6000, but it can do 6000C28 100% stable. The mainstream desktop platform is able to run 6400+ only at command rate 2T, so 6000C28 at 1T is actually a very good OC. Since these are R-DIMMs, the memory controller voltage makes no difference. The registered buffer on the DIMM takes the load off the memory controller. You can run stock memory controller voltage when overclocking for lower CPU temps. The register also adds about 30ns of latency compared to mainstream desktop.
ASRock sent me several beta BIOSes which fixed most of the above issues. They released a fix for each issue within 1 or 2 days of me creating a support ticket. I am shocked at the excellent support.
The ASrock + a W5-2445 would interest me for a general workstation with ECC DDR5 + pcie5. I'm not interested in overclocking anything, just want a decent modern desktop setup. But roughly that's still $880 + $840 = $1720.. Not even close to my budget. Maybe just wait and look at what comes after AL/RL + W680 instead..
--
if you have EPC621D8A and don't ask for Beta BIOS you stuck at BIOS 2.10 from 4/25/2019 and this is shocking me.
nice for you.
if you have EPC621D8A and don't ask for Beta BIOS you stuck at BIOS 2.10 from 4/25/2019 and this is shocking me.
Also shocking is that no mobo manufacturer is interested in switching to some project like coreboot or whatever there is.. I'm sick and tired of this issue with proprietary bios code.
they can not use copyright protected code.
AMI BIOS is not proprietary. but the used code for CPU initializing (Intel RC/MRC, AMD AGESA) must be.
let's have AI design an opensource CPU and crowdfunding to manufacture it in some indian sweatshop. Let's break free!
edit: Chinese Researchers Used AI to Design RISC-V CPU in Under 5 Hours
Bug Forces Intel to Halt Some Xeon Sapphire Rapids Shipments
"We became aware of an issue on a subset of 4th Generation Intel Xeon Medium Core Count Processors (SPR-MCC) that could interrupt system operation under certain conditions and are actively investigating. This issue was not observed when running commercially available software, and other 4th Generation Intel Xeon processor SKUs (i.e., XCC and HBM) have not exhibited the issue. Out of an abundance of caution, we did temporarily pause some SPR MCC shipments while we gained confidence in the expected firmware mitigation and expect to release remaining shipments shortly." — Intel Spokesperson to Tom's Hardware."
edit: Chinese Researchers Used AI to Design RISC-V CPU in Under 5 Hours
Bug Forces Intel to Halt Some Xeon Sapphire Rapids Shipments
"We became aware of an issue on a subset of 4th Generation Intel Xeon Medium Core Count Processors (SPR-MCC) that could interrupt system operation under certain conditions and are actively investigating. This issue was not observed when running commercially available software, and other 4th Generation Intel Xeon processor SKUs (i.e., XCC and HBM) have not exhibited the issue. Out of an abundance of caution, we did temporarily pause some SPR MCC shipments while we gained confidence in the expected firmware mitigation and expect to release remaining shipments shortly." — Intel Spokesperson to Tom's Hardware."
Last edited: