Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
moondev
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
19 ms
·
241.
▲
by
moondev
3y ago
Can't have the Apple car shipping before the Apple 5g modem.
242.
▲
by
moondev
3y ago
Don't you need CPUs, PCIE lanes, NVME and cooling systems to deploy GPGPU(s) and DPU(s) - or are you suggesting hopper will replace it?
243.
▲
by
moondev
3y ago
> a mid range server CPU of today will be about as fast as an entry model CPU next year What about the core count and memory capacity
244.
▲
by
moondev
3y ago
On Linux you can specify any cloud-init qcow2 image with file:// or http(s):// and it works beautifully
245.
▲
by
moondev
3y ago
Behold, an inline USB QMK controller. Just carry one on your keychain https://yaowei.dev/posts/usb-to-usb-converter/
246.
▲
by
moondev
3y ago
Can you? Pull through cache is awesome but I have been leaning on harbor registry to implement it. Harbor is great but is a full featured registry application. Would be sweet if pull through can be quickly spun up ad hoc.
247.
▲
by
moondev
3y ago
Spiritual successor to https://github.com/gravitational/gravity ?
248.
▲
by
moondev
3y ago
https://github.com/AmpereComputing/Windows-11-On-Ampere
249.
▲
by
moondev
3y ago
Running windows for arm requires standardized EFI firmware https://www.arm.com/architecture/system-architectures/system... Many arm64 devices run windows for arm today, including non Qualcomm devices like ampere d
250.
▲
by
moondev
3y ago
https://www.tomshardware.com/news/ampere-64-core-arm-worksta...
251.
▲
by
moondev
3y ago
A100 has no display outputs or integrated fan which makes it a poor fit for a workstation.
252.
▲
by
moondev
3y ago
https://kubernetes.io/docs/concepts/storage/volume-pvc-datas...
253.
▲
by
moondev
3y ago
Another alternative, launch a kvm with GPU passthrough and use cloudinit to launch sunshine and the game, or just use the monitor directly. https://kubevirt.io/user-guide/virtual_machines/host-devices... Declarati
254.
▲
by
moondev
3y ago
I'm curious if an 8GB MacBook can run a macOS VM, and if so how much memory can be allocated to it.
255.
▲
by
moondev
3y ago
Now we just need a kubernetes cluster with coredns to deploy eureka! http://eureka-0.eureka.default.svc.cluster.local:eureka:8761/eureka http://eureka-1.eureka.default.svc.cluster.local:eureka:8761/eu
256.
▲
by
moondev
3y ago
Some motherboards and pci slots have various methods of "locking" a card in place. Seems like this may have been forgotten and force was used to remove the card. I don't see how the weight of the card could crack the PCB, the
257.
▲
by
moondev
3y ago
I really like the "tab overview" interface and animations. Very slick and functional.
258.
▲
by
moondev
3y ago
Kustomize also supports helm charts as a "resource" which makes it handy to do last mile modifications of values and 'non value exposed" items without touching or forking the upstream chart.
259.
▲
by
moondev
3y ago
With that logic why does it even have a gigabit uplink, after all three are still people out there that are not "hitting their 100Mbit/s" yet?
260.
▲
by
moondev
3y ago
> It shows. Mission accomplished Fusion now supports running Arm virtual machines on Apple Silicon Macs https://docs.vmware.com/en/VMware-Fusion/13.0/rn/vmware-fusi...
261.
▲
by
moondev
3y ago
Say I want to launch a vm with 8GB reserved memory. On a 10-12 GB machine I would have 2 to 4GB of memory left to actually allocate to the system. I don't see how unified memory magically fixes this. It's actually worse because e
262.
▲
by
moondev
3y ago
Since 8GB unified memory is "equivalent to a 16GB non Mac" Is 128GB unified memory MacBook basically a 256GB Nvidia GPU(s)?
263.
▲
by
moondev
3y ago
I can recommend NIZ keyboards if you are looking for something totally different. The switches are electro capacitive dome not mechanical. Very light and effortless to type and very quiet as well with a hint of tactileness. They are expensi
264.
▲
by
moondev
3y ago
Adding .pibb to a gist url is another nice trick
265.
▲
by
moondev
3y ago
Framework mainboards are all usb-c
266.
▲
by
moondev
3y ago
Since bhyve can't do nested virtualization, i guess they can't "dog food" the platform during development. I wonder if for that they use Linux. The same hardware and software using kvm would be much more appealing. Not b
267.
▲
by
moondev
3y ago
Has anyone built or attempted to build their own flash array, as in 20+ nvme u.2 disks in a hot swap chassis and consume via nvme-of or nvme-tcp I see there are tri-mode backplanes but looking for something that skips the hba and connects v
268.
▲
by
moondev
3y ago
I have an AGX Xavier running ESXi arm edition. This makes it easy to run any Linux aarch64 distro that provides a uefi iso installer. It even supports running windows for arm vms. Great way to put the 32G of memory and NVME to use
269.
▲
by
moondev
3y ago
Needing to convert yaml to hcl seems like a burden. A killer feature of HCL would be if it supported syntax highlighted yaml. Or if the HCL manifest was in json format, it perhaps could be easily then converted to yaml as the "human re
270.
▲
by
moondev
3y ago
A note for those unaware, consumer grade NVME devices (basically all m.2 formfactor drives from my experience) only support a single namespace. If you want to explore creating multiple namespaces you will need an enterprise grade u.2 drive.
More ›