Search results

The Commiecast customers are all offline and hence can't complain on the Internet.

I have the same monitor and no problems. IMHO it is a good unit. I spend more and more time on my Macbook, though.

A couple of nails would do the job just fine :D

I still have a DFI SLI-DR with a CABNE Opteron socket 939 in my retro gaming machine. A board that was a lot of fun. DFI also made a cheap line of boards. IIRC they were no good. DFI went to make industrial computer components, dunno whether they are still around.

I dunno. I would at best put a Steam library on one of these, and only if the games don't have investments in mods. But for games that can easily be restored and not constantly reinstalled they probably rock.

I second that, I had problems with more expensive Asus boards, but the cheaper ones have (as far as I remember) always worked for me. Right now I like their Prime sub-brand.

Right now on Ebay you can pick up 128 GB (4x 32) of DDR4 registered for $110 or so. UDIMMs are about 3,5x more :sad:

I have 2 Gigabyte AM4 systems running with unbuffered DDR4 ECC. I know it works because I got some bad modules and they reported into the OS just fine. But again, this is limited to 128 GB of RAM.

But then you are subject to the usual desktop limitations wrt maximum amount of RAM. The whole point about registered RAM is that you can have more of it.

That one doesn't take registered RAM, only unbuffered. It will probably do ECC. I am not aware of small boards taking registered DDR3. For DDR4 there are various Xeon-D options.

Good, so the 4 bad ones are definitely fixed by the March update. Now I just have to hope T-mobile doesn't want to attack me :)

I think they have been doing staggered updates for a while now. I always have to get the update by asking for it (on T-mobile/US).

Not anymore with DDR5.

The interesting thing about the 16" modular bay for the GPU is that you can also get a module that adds another two M.2 SSD slots. On a related note, I hope there is no Optimus like switching of GPUs in play when you use a GPU module.

Well, one reason for waiting is slot space. The 4090 is 3.5 slots wide. In the future you'd get the same graphics power in 2 slots wide.

My Pixel 6 did get the March update after much delay. But I have not seen direct, clear confirmation that it fixes this vulnerability.

I had broken board by all three except Gigabyte, but there were BIOS problems on the GB. I went back to buying Asus, but I would not RMA a board, I'd just get a new one. So I buy their medium priced one, e.g. Asus Prime. Having said that, most of my stuff except my gaming machine is...

Who bought the Egg?

NFS support Unix/POSIX semantics much better. Whether it is faster depends on what you are currently doing, but generally NFS will easily max out 1 GbE connections.

Even if one pass of zeros doesn't do it, another pass with 1s would cause much more damage, no?

They say while posting the usual screenshot with no UI...

The root of the problem is that the manufacturers don't give specific information that could be used to estimate the durability. Especially for flash cards and USB sticks.

That's probably it. Doesn't work for me.

So this is the patch that fixes this week's panic about WiFi calling, right? Right?

T-mobile has it for me now.

Yeah. Pull the band-aid and change rather sooner than later.

Yeah, the situation with ECC UDIMM is ridiculous, both on DDR4 and DDR5. ECC registered DDR4 can be had for a fraction of the price. I am almost not angry that current Threadrippers are only available as Pros (Pro takes registered RAM). The vendor locks make me very sad, though. I find Golden...

What exactly does that do?

Critical vulnerabilities. Turn wifi calling off. I still can't get a patch for my Pixel 6 (not to mention it is unclear whether the March patch addresses these issues): https://arstechnica.com/information-technology/2023/03/critical-vulnerabilities-allow-some-android-phones-to-be-hacked/

Reading only one mirror in a RAID1 allows the drives to operate in parallel on concurrent reads. That is a speedup of 2 for 2 disks, 3 for 3 disks etc on random block reads. As discussed in this thread, it also saves bus bandwidth in software RAID.

The same way as it does with single disk: if the drive doesn't report it the corrupt data goes through to the user. As I said, it is a choice.

That is not necessarily true. In RAID1 you only have to read one of the drives, reading all and comparing is optional for reads. This also speeds up random, concurrent seeks by a factor of how many drives you have in your raid1. Likewise, raid5 data can be read without one drive if you so wish...

Just paid for 256 GB for my server upgrade. Can't change vote.

Had anybody seen prices on DDR5 registered memory yet?

Just waiting for Apple's switch to RISC-V. I am actually sure it will come. Might take a long while, but it will come. Depending on companies that can go crazy at any point in time (combined with an acquisition or not) is just not a winning business model anymore.

A dedicated server build might have different scheduler or preemption model than the same vendor's desktop build. But that is easy to change after install and it will be far from clear what is better in what situation.

Apologies, I thought the MX500 is the higher tier.

The MX500 4 TB is only $240. Probably a better buy: https://www.amazon.com/gp/product/B09FRRWVWX/ref=ox_sc_act_title_1?smid=ATVPDKIKX0DER&psc=1

I hear there is a March patch with important security updates, but my P6 on T-mobile doesn't seem to be able to get it OTA.

What's that telemetry plugin sending, anyway?