Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Need advice on pay-as-you-go CPU Servers

Hi guys:
I will do some NLP work and need to training my models, could you help advice me some providers on pay-as-you-go CPU Servers? Or do you know some platforms serve clients like me? No GPU requirement.

Maybe OVH is a choice, if any others?

Thank you for your help

Comments

  • DPDP Administrator, The Domain Guy

    Vultr?

    Thanked by 1Creling
  • JordJord Moderator, Host Rep, Veteran, Megathread Squad

    Did anyone mention Hetzner?

    Thanked by 2Creling Hetzner_OL
  • Hetzner and Lunanode (dedicated vCPU/compute optimized) has per hour billing too

  • @Jord said:
    Did anyone mention Hetzner?

    Hetzner

    Thanked by 2Creling Hetzner_OL
  • Hetzner of course

    Thanked by 2ITLabs Hetzner_OL
  • Hetzner

    Thanked by 1Hetzner_OL
  • dahartigandahartigan Member
    edited November 2019

    On a more serious note, LunaNode is excellent value and their performance and stability is even better. They have Canada and France options. $USD 3.50/mo gets you started.

    BinaryLane for Brisbane, Sydney, Melbourne and Perth in Australia. Unbeatable low-end value, again, excellent performance and stability. $AUD 4/mo gets you a seat on a server in Brisbane. Blows my mind.

    I would recommend both those providers all day long. My go-to for "I really need this shit to work and not worry about it going offline" projects.

    Thanked by 1termau
  • @Creling said:
    Hi guys:
    I will do some NLP work and need to training my models, could you help advice me some providers on pay-as-you-go CPU Servers? Or do you know some platforms serve clients like me? No GPU requirement.

    Maybe OVH is a choice, if any others?

    Thank you for your help

    Do you have a large dataset? If you do, you might take a really long time without a dedicated GPU, and I am not sure if cloud providers have any limits on CPU usage.

  • @poisson said:

    @Creling said:
    Hi guys:
    I will do some NLP work and need to training my models, could you help advice me some providers on pay-as-you-go CPU Servers? Or do you know some platforms serve clients like me? No GPU requirement.

    Maybe OVH is a choice, if any others?

    Thank you for your help

    Do you have a large dataset? If you do, you might take a really long time without a dedicated GPU, and I am not sure if cloud providers have any limits on CPU usage.

    LunaNode gives you the option to pay extra for dedicated use of CPU, it's actually kinda neat. I did that so I could transcode some large videos without abusing the node, and it worked out nicely. Not even expensive either.

    Thanked by 2uptime poisson
  • CrelingCreling Member
    edited November 2019

    dahartigan said: LunaNode gives you the option to pay extra for dedicated use of CPU, it's actually kinda neat. I did that so I could transcode some large videos without abusing the node, and it worked out nicely. Not even expensive either.

    thanks, maybe I will try LonaNode Compute-Optimized Virtual Machines, btw, do you know which cpu they use?

  • @Jord said:
    Did anyone mention Hetzner?

    ah, yes, I happen to have a Hetzner account 😄

  • poisson said: Do you have a large dataset? If you do, you might take a really long time without a dedicated GPU, and I am not sure if cloud providers have any limits on CPU usage.

    ah, some effective algorithm like gensim will not cost GPU resource 😜. No GPU is not to purse cheap, but to fit my need. 😝

  • @Creling said:

    dahartigan said: LunaNode gives you the option to pay extra for dedicated use of CPU, it's actually kinda neat. I did that so I could transcode some large videos without abusing the node, and it worked out nicely. Not even expensive either.

    thanks, maybe I will try LonaNode Compute-Optimized Virtual Machines, btw, do you know which cpu they use?

    $ cat /proc/cpuinfo
    processor       : 0
    vendor_id       : GenuineIntel
    cpu family      : 6
    model           : 62
    model name      : Intel(R) Xeon(R) CPU E5-2690 v2 @ 3.00GHz
    stepping        : 4
    microcode       : 0x1
    cpu MHz         : 2999.998
    cache size      : 25600 KB
    physical id     : 0
    siblings        : 1
    core id         : 0
    cpu cores       : 1
    apicid          : 0
    initial apicid  : 0
    fpu             : yes
    fpu_exception   : yes
    cpuid level     : 13
    wp              : yes
    flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon rep_good nopl cpuid tsc_known_freq pni pclmulqdq vmx ssse3 cx16 pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm pti tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust smep erms xsaveopt
    bugs            : cpu_meltdown spectre_v1 spectre_v2 spec_store_bypass l1tf mds swapgs
    bogomips        : 5999.99
    clflush size    : 64
    cache_alignment : 64
    address sizes   : 40 bits physical, 48 bits virtual
    power management:
    

    I hope that helps. Not sure if their entire fleet is identical though. Maybe @perennate can shed some light :-)

    Thanked by 1uptime
  • @Creling said:

    poisson said: Do you have a large dataset? If you do, you might take a really long time without a dedicated GPU, and I am not sure if cloud providers have any limits on CPU usage.

    ah, some effective algorithm like gensim will not cost GPU resource 😜. No GPU is not to purse cheap, but to fit my need. 😝

    Oh. I will check out gensim. Wasn't aware of non-GPU based NLP toolkits; I am too used to using GPU-based ones.

  • @Creling said:
    Hi guys:
    I will do some NLP work and need to training my models, could you help advice me some providers on pay-as-you-go CPU Servers? Or do you know some platforms serve clients like me? No GPU requirement.

    Maybe OVH is a choice, if any others?

    Thank you for your help

    for hourly bare metal dedicated with exotic network and gpu configs check out packet.net https://www.packet.com/cloud/servers/

    Thanked by 1uptime
  • eva2000 said: for hourly bare metal dedicated with exotic network and gpu configs check out packet.net https://www.packet.com/cloud/servers/

    happy to see such a platform, I will mark it.

  • CrelingCreling Member
    edited November 2019

    poisson said: Oh. I will check out gensim. Wasn't aware of non-GPU based NLP toolkits; I am too used to using GPU-based ones.

    maybe just in some special cases, I will do some word2vec.

    Thanked by 1poisson
  • Scaleway have baremetal server with per hour pricing so you can rent whole powerfull server for 1 hour for cheap.
    On other hand if you plan to do alot stuff monthly renting server may be cheaper option.
    Scaleway for example have server for around 329 euro a month (in hourly billing). You could get similar server when it comes to CPU foraround $100 a month. Sometines even cheaper.

    hetzner cloud is also good when it comes to prive/performance ratio

  • @Creling if you want free Upcloud credits I can give you some to fire up their 20 core servers for 1.5 days

    Thanked by 1hjlow
  • You can try hetzner or linode (only if you need asian location)

    Thanked by 1Creling
  • poisson said: @Creling if you want free Upcloud credits I can give you some to fire up their 20 core servers for 1.5 days

    hhhh, 0.5 day will be enough for trialing. I might pm you when I need it. thank you 2333

  • Upcloud, but they asked for CC/DC

  • Wholesaleinternet provide hourly charged servers, the last one that I saw started from like 7 cents, this was yesterday though and dont know if its still there.

Sign In or Register to comment.