Anyone tried the new AMD Vultr Instances?

Comments

  • Just looked at the prices and for the AMD instances they start at 28 USD per instance which is a bit over LES levels.

  • kuroneko@test1:~$ bash index.html
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2022-02-18                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Wed 16 Mar 2022 01:19:51 PM UTC
    
    Basic System Information:
    ---------------------------------
    Processor  : AMD EPYC-Rome Processor
    CPU cores  : 1 @ 1996.249 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ❌ Disabled
    RAM        : 1.9 GiB
    Swap       : 0.0 KiB
    Disk       : 23.4 GiB
    
    fio Disk Speed Tests (Mixed R/W 50/50):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 12.00 MB/s    (3.0k) | 196.28 MB/s   (3.0k)
    Write      | 12.00 MB/s    (3.0k) | 197.32 MB/s   (3.0k)
    Total      | 24.01 MB/s    (6.0k) | 393.60 MB/s   (6.1k)
               |                      |
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 121.75 MB/s    (237) | 198.89 MB/s    (194)
    Write      | 128.22 MB/s    (250) | 212.14 MB/s    (207)
    Total      | 249.98 MB/s    (487) | 411.04 MB/s    (401)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed
                    |                           |                 |
    Clouvider       | London, UK (10G)          | 179 Mbits/sec   | 794 Mbits/sec
    Online.net      | Paris, FR (10G)           | 53.6 Mbits/sec  | 586 Mbits/sec
    WorldStream     | The Netherlands (10G)     | 1.00 Gbits/sec  | busy
    WebHorizon      | Singapore (400M)          | 771 Mbits/sec   | 455 Mbits/sec
    Clouvider       | NYC, NY, US (10G)         | 513 Mbits/sec   | 632 Mbits/sec
    Velocity Online | Tallahassee, FL, US (10G) | 837 Mbits/sec   | 1.69 Gbits/sec
    Clouvider       | Los Angeles, CA, US (10G) | 1.05 Gbits/sec  | 882 Mbits/sec
    Iveloz Telecom  | Sao Paulo, BR (2G)        | busy            | 498 Mbits/sec
    
    Geekbench 5 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 947
    Multi Core      | 941
    Full Test       | https://browser.geekbench.com/v5/cpu/13582551
    
    Thanked by (3)ehab Aidan FrankZ
  • @serverollie said:
    Just looked at the prices and for the AMD instances they start at 28 USD per instance which is a bit over LES levels.

    Vultr is good only for hourly instance. I used to use it as a game server and delete it when done. Cost around $1-2 a month.

  • @kuroneko23 said:

    kuroneko@test1:~$ bash index.html
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2022-02-18                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Wed 16 Mar 2022 01:19:51 PM UTC
    
    Basic System Information:
    ---------------------------------
    Processor  : AMD EPYC-Rome Processor
    CPU cores  : 1 @ 1996.249 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ❌ Disabled
    RAM        : 1.9 GiB
    Swap       : 0.0 KiB
    Disk       : 23.4 GiB
    
    fio Disk Speed Tests (Mixed R/W 50/50):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 12.00 MB/s    (3.0k) | 196.28 MB/s   (3.0k)
    Write      | 12.00 MB/s    (3.0k) | 197.32 MB/s   (3.0k)
    Total      | 24.01 MB/s    (6.0k) | 393.60 MB/s   (6.1k)
               |                      |
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 121.75 MB/s    (237) | 198.89 MB/s    (194)
    Write      | 128.22 MB/s    (250) | 212.14 MB/s    (207)
    Total      | 249.98 MB/s    (487) | 411.04 MB/s    (401)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed
                    |                           |                 |
    Clouvider       | London, UK (10G)          | 179 Mbits/sec   | 794 Mbits/sec
    Online.net      | Paris, FR (10G)           | 53.6 Mbits/sec  | 586 Mbits/sec
    WorldStream     | The Netherlands (10G)     | 1.00 Gbits/sec  | busy
    WebHorizon      | Singapore (400M)          | 771 Mbits/sec   | 455 Mbits/sec
    Clouvider       | NYC, NY, US (10G)         | 513 Mbits/sec   | 632 Mbits/sec
    Velocity Online | Tallahassee, FL, US (10G) | 837 Mbits/sec   | 1.69 Gbits/sec
    Clouvider       | Los Angeles, CA, US (10G) | 1.05 Gbits/sec  | 882 Mbits/sec
    Iveloz Telecom  | Sao Paulo, BR (2G)        | busy            | 498 Mbits/sec
    
    Geekbench 5 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 947
    Multi Core      | 941
    Full Test       | https://browser.geekbench.com/v5/cpu/13582551
    

    Disappointing.

  • A big change for Vultr. They are now offering segmented resources system rather than the locked-in tiered, which is good becuase if you needed more ram but not CPU it was too bad

  • Woah, that's a low Geekbench.

    I saw the announcement earlier today, and was considering deploying a new instance to replace an old one, but that's kinda looking low where I don't find it so attractive compared to their old High Frequency plans.

    One can hope it's just that their nodes are overloaded, but then it's odd coming from a company such as Vultr. Might give it a try later.

  • edited March 2022

    For comparison, this is the result of my current instance, which is part of the old High Frequency plans:

    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2022-02-18                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Wed Mar 16 21:00:47 -04 2022
    
    Basic System Information:
    ---------------------------------
    Processor  : Intel Core Processor (Skylake, IBRS)
    CPU cores  : 1 @ 3792.000 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ❌ Disabled
    RAM        : 1.9 GiB
    Swap       : 997.6 MiB
    Disk       : 63.0 GiB
    
    fio Disk Speed Tests (Mixed R/W 50/50):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 291.45 MB/s  (72.8k) | 633.06 MB/s   (9.8k)
    Write      | 292.22 MB/s  (73.0k) | 636.39 MB/s   (9.9k)
    Total      | 583.67 MB/s (145.9k) | 1.26 GB/s    (19.8k)
               |                      |                     
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 699.37 MB/s   (1.3k) | 740.17 MB/s    (722)
    Write      | 736.53 MB/s   (1.4k) | 789.47 MB/s    (770)
    Total      | 1.43 GB/s     (2.8k) | 1.52 GB/s     (1.4k)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed     
                    |                           |                 |                
    Clouvider       | London, UK (10G)          | 1.82 Gbits/sec  | 1.60 Gbits/sec 
    Online.net      | Paris, FR (10G)           | 2.01 Gbits/sec  | 1.81 Gbits/sec 
    WorldStream     | The Netherlands (10G)     | busy            | busy           
    WebHorizon      | Singapore (400M)          | 567 Mbits/sec   | 409 Mbits/sec  
    Clouvider       | NYC, NY, US (10G)         | 3.58 Gbits/sec  | 1.59 Gbits/sec 
    Velocity Online | Tallahassee, FL, US (10G) | 4.01 Gbits/sec  | 3.03 Gbits/sec 
    Clouvider       | Los Angeles, CA, US (10G) | 2.73 Gbits/sec  | 689 Mbits/sec  
    Iveloz Telecom  | Sao Paulo, BR (2G)        | 1.42 Gbits/sec  | 1.69 Gbits/sec 
    
    iperf3 Network Speed Tests (IPv6):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed     
                    |                           |                 |                
    Clouvider       | London, UK (10G)          | 1.86 Gbits/sec  | 1.75 Gbits/sec 
    Online.net      | Paris, FR (10G)           | 2.01 Gbits/sec  | busy           
    WorldStream     | The Netherlands (10G)     | 1.46 Gbits/sec  | busy           
    WebHorizon      | Singapore (400M)          | 598 Mbits/sec   | 381 Mbits/sec  
    Clouvider       | NYC, NY, US (10G)         | 6.47 Gbits/sec  | 1.38 Gbits/sec 
    Clouvider       | Los Angeles, CA, US (10G) | 2.82 Gbits/sec  | 2.65 Gbits/sec 
    
    Geekbench 5 Benchmark Test:
    ---------------------------------
    Test            | Value                         
                    |                               
    Single Core     | 964                           
    Multi Core      | 956                           
    Full Test       | https://browser.geekbench.com/v5/cpu/13595963
    

    EDIT: I just noticed, are new instances now being throttled to 1 Gbit instead of 10 Gbit shared?

  • @serverollie said:
    Just looked at the prices and for the AMD instances they start at 28 USD per instance which is a bit over LES levels.

    Also, this is for the dedicated resources. The new "Cloud Compute" has both plans for what were Cloud Compute ($5/mo $10/mo etc.) and High Frequency ($6/mo $12/mo etc.) plans.

  • They got rid of the high freq $6/m plan? I used that a little and it was great, oh well. For $6/m now it's harder to compete with Hetzner.

  • cybertechcybertech OGBenchmark King

    looks weak.

    I bench YABS 24/7/365 unless it's a leap year.

  • edited March 2022

    @willie said:
    They got rid of the high freq $6/m plan? I used that a little and it was great, oh well. For $6/m now it's harder to compete with Hetzner.

    Again:

    @JeDaYoshi said:
    The new "Cloud Compute" has both plans for what were Cloud Compute ($5/mo $10/mo etc.) and High Frequency ($6/mo $12/mo etc.) plans.

    They are divided in each category, of older and newer processors.

  • Oh I see, the high frequency plans are still there at the same prices as before, and hopefully the same overall specs. Sounds good. I see they now have a $2.50/month 0.5gb plan. Maybe I can switch my $5 plan to that, for my super low traffic nextcloud. Thanks.

  • edited March 2022

    @JeDaYoshi said:
    EDIT: I just noticed, are new instances now being throttled to 1 Gbit instead of 10 Gbit shared?

    Does the location matters? Because I deployed it in Singapore region.

  • @kuroneko23 said:

    @JeDaYoshi said:
    EDIT: I just noticed, are new instances now being throttled to 1 Gbit instead of 10 Gbit shared?

    Does the location matters? Because I deployed it in Singapore region.

    Actually I am dumb. I didn't notice that you had gotten a 1.69 Gbits/sec in Florida.
    My first YABS is from Miami, however, this one is from my old High Frequency Singapore instance:

    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2022-02-18                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Thu 17 Mar 2022 04:45:05 AM UTC
    
    Basic System Information:
    ---------------------------------
    Processor  : Intel Core Processor (Skylake, IBRS)
    CPU cores  : 1 @ 3407.998 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ❌ Disabled
    RAM        : 976.4 MiB
    Swap       : 488.2 MiB
    Disk       : 31.8 GiB
    
    fio Disk Speed Tests (Mixed R/W 50/50):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 254.32 MB/s  (63.5k) | 976.03 MB/s  (15.2k)
    Write      | 254.99 MB/s  (63.7k) | 981.17 MB/s  (15.3k)
    Total      | 509.32 MB/s (127.3k) | 1.95 GB/s    (30.5k)
               |                      |                     
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 1.06 GB/s     (2.0k) | 1.10 GB/s     (1.0k)
    Write      | 1.12 GB/s     (2.1k) | 1.17 GB/s     (1.1k)
    Total      | 2.19 GB/s     (4.2k) | 2.27 GB/s     (2.2k)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed     
                    |                           |                 |                
    Clouvider       | London, UK (10G)          | 994 Mbits/sec   | 1.03 Gbits/sec 
    Online.net      | Paris, FR (10G)           | 763 Mbits/sec   | busy           
    WorldStream     | The Netherlands (10G)     | 952 Mbits/sec   | 1.05 Gbits/sec 
    WebHorizon      | Singapore (400M)          | 941 Mbits/sec   | 496 Mbits/sec  
    Clouvider       | NYC, NY, US (10G)         | 712 Mbits/sec   | 702 Mbits/sec  
    Velocity Online | Tallahassee, FL, US (10G) | 751 Mbits/sec   | 720 Mbits/sec  
    Clouvider       | Los Angeles, CA, US (10G) | 982 Mbits/sec   | 993 Mbits/sec  
    Iveloz Telecom  | Sao Paulo, BR (2G)        | busy            | 442 Mbits/sec  
    
    iperf3 Network Speed Tests (IPv6):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed     
                    |                           |                 |                
    Clouvider       | London, UK (10G)          | 1.08 Gbits/sec  | 1.04 Gbits/sec 
    Online.net      | Paris, FR (10G)           | 717 Mbits/sec   | busy           
    WorldStream     | The Netherlands (10G)     | busy            | 1.01 Gbits/sec 
    WebHorizon      | Singapore (400M)          | 924 Mbits/sec   | 482 Mbits/sec  
    Clouvider       | NYC, NY, US (10G)         | 755 Mbits/sec   | 742 Mbits/sec  
    Clouvider       | Los Angeles, CA, US (10G) | 1.04 Gbits/sec  | 994 Mbits/sec  
    
    Geekbench 5 Benchmark Test:
    ---------------------------------
    Test            | Value                         
                    |                               
    Single Core     | 1073                          
    Multi Core      | 1024                          
    Full Test       | https://browser.geekbench.com/v5/cpu/13599611
    

    Seems like about the same average, although I swear I had gotten better results before.
    But at least that's a relief, got scared for a moment. Excuse me.
    I wonder if results are somehow different in CPU score etc. now, but considering CPU clock speed is lower... Yeah. It's disappointing.


  • AMD Optimized Cloud Compute (Dedicated vCPU) LA
    kuroneko@yabs1:~$ bash index.html
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2022-02-18                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Thu 17 Mar 2022 05:12:42 AM UTC
    
    Basic System Information:
    ---------------------------------
    Processor  : AMD EPYC-Rome Processor
    CPU cores  : 1 @ 1996.249 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ❌ Disabled
    RAM        : 1.9 GiB
    Swap       : 0.0 KiB
    Disk       : 23.4 GiB
    
    fio Disk Speed Tests (Mixed R/W 50/50):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 61.00 MB/s   (15.2k) | 232.76 MB/s   (3.6k)
    Write      | 61.12 MB/s   (15.2k) | 233.99 MB/s   (3.6k)
    Total      | 122.12 MB/s  (30.5k) | 466.75 MB/s   (7.2k)
               |                      |
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 227.31 MB/s    (443) | 398.42 MB/s    (389)
    Write      | 239.39 MB/s    (467) | 424.95 MB/s    (414)
    Total      | 466.70 MB/s    (910) | 823.38 MB/s    (803)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed
                    |                           |                 |
    Clouvider       | London, UK (10G)          | 187 Mbits/sec   | 1.07 Gbits/sec
    Online.net      | Paris, FR (10G)           | busy            | busy
    WorldStream     | The Netherlands (10G)     | 1.08 Gbits/sec  | 792 Mbits/sec
    WebHorizon      | Singapore (400M)          | 641 Mbits/sec   | 399 Mbits/sec
    Clouvider       | NYC, NY, US (10G)         | 1.54 Gbits/sec  | 1.83 Gbits/sec
    Velocity Online | Tallahassee, FL, US (10G) | 1.40 Gbits/sec  | 2.70 Gbits/sec
    Clouvider       | Los Angeles, CA, US (10G) | 2.55 Gbits/sec  | 3.33 Gbits/sec
    Iveloz Telecom  | Sao Paulo, BR (2G)        | 877 Mbits/sec   | 1.19 Gbits/sec
    
    iperf3 Network Speed Tests (IPv6):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed
                    |                           |                 |
    Clouvider       | London, UK (10G)          | 1.39 Gbits/sec  | 1.07 Gbits/sec
    Online.net      | Paris, FR (10G)           | 21.5 Mbits/sec  | 66.1 Mbits/sec
    WorldStream     | The Netherlands (10G)     | 1.16 Gbits/sec  | 1.12 Gbits/sec
    WebHorizon      | Singapore (400M)          | 756 Mbits/sec   | 430 Mbits/sec
    Clouvider       | NYC, NY, US (10G)         | 3.05 Gbits/sec  | 2.60 Gbits/sec
    Clouvider       | Los Angeles, CA, US (10G) | 2.51 Gbits/sec  | 3.75 Gbits/sec
    
    Geekbench 5 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 951
    Multi Core      | 953
    Full Test       | https://browser.geekbench.com/v5/cpu/13600000
    



    AMD Cloud Compute (Shared vCPU) LA
    kuroneko@yabs1:~$ bash index.html -i
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2022-02-18                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Thu 17 Mar 2022 05:30:07 AM UTC
    
    Basic System Information:
    ---------------------------------
    Processor  : AMD EPYC-Rome Processor
    CPU cores  : 1 @ 1996.249 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ❌ Disabled
    RAM        : 1.9 GiB
    Swap       : 0.0 KiB
    Disk       : 46.9 GiB
    
    fio Disk Speed Tests (Mixed R/W 50/50):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 39.33 MB/s    (9.8k) | 128.54 MB/s   (2.0k)
    Write      | 39.41 MB/s    (9.8k) | 129.21 MB/s   (2.0k)
    Total      | 78.74 MB/s   (19.6k) | 257.76 MB/s   (4.0k)
               |                      |
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 146.10 MB/s    (285) | 234.30 MB/s    (228)
    Write      | 153.87 MB/s    (300) | 249.91 MB/s    (244)
    Total      | 299.97 MB/s    (585) | 484.21 MB/s    (472)
    
    Geekbench 5 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 950
    Multi Core      | 938
    Full Test       | https://browser.geekbench.com/v5/cpu/13600201
    


    I guess the processor information in YABS is incorrect maybe because no host-passthrough? It should be AMD Epyc Milan, probably 7713.
    Thanked by (2)Aidan JeDaYoshi
  • edited March 2022

    So the new AMD plans are pretty disappointing for the price and the hype, especially since the older Intel CPUs beat them (albeit by a little, but still), and even the NVMe drives are better than what they threw on the newer nodes. And I bet they're still not RAID'd.

    Noted.

    I still don't understand why the clock speed is so low, though.

  • Somehow I got higher single core score on 2c4g


    Vultr Cloud Compute SG ($0.06/h)
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2022-02-18                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Thu 17 Mar 2022 06:14:51 AM UTC
    
    Basic System Information:
    ---------------------------------
    Processor  : AMD EPYC-Rome Processor
    CPU cores  : 2 @ 1996.249 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ❌ Disabled
    RAM        : 3.8 GiB
    Swap       : 7.8 GiB
    Disk       : 93.8 GiB
    
    fio Disk Speed Tests (Mixed R/W 50/50):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 19.68 MB/s    (4.9k) | 239.07 MB/s   (3.7k)
    Write      | 19.68 MB/s    (4.9k) | 240.33 MB/s   (3.7k)
    Total      | 39.37 MB/s    (9.8k) | 479.40 MB/s   (7.4k)
               |                      |
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 96.02 MB/s     (187) | 122.60 MB/s    (119)
    Write      | 101.12 MB/s    (197) | 130.76 MB/s    (127)
    Total      | 197.14 MB/s    (384) | 253.37 MB/s    (246)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed
                    |                           |                 |
    Clouvider       | London, UK (10G)          | 1.12 Gbits/sec  | 908 Mbits/sec
    Online.net      | Paris, FR (10G)           | 4.37 Gbits/sec  | 2.89 Gbits/sec
    Clouvider       | NYC, NY, US (10G)         | 754 Mbits/sec   | 760 Mbits/sec
    
    iperf3 Network Speed Tests (IPv6):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed
                    |                           |                 |
    Clouvider       | London, UK (10G)          | 1.07 Gbits/sec  | 1.06 Gbits/sec
    Online.net      | Paris, FR (10G)           | 1.77 Gbits/sec  | busy
    Clouvider       | NYC, NY, US (10G)         | 749 Mbits/sec   | 757 Mbits/sec
    
    Geekbench 5 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 1104
    Multi Core      | 2185
    Full Test       | https://browser.geekbench.com/v5/cpu/13600987
    



    AWS EC2 c5a.large SG ($0.0311 for spot pricing, $0.088 for normal pricing)
    ubuntu@ip-172-31-9-125:~$ bash index.html -r
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2022-02-18                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Thu Mar 17 06:20:38 UTC 2022
    
    Basic System Information:
    ---------------------------------
    Processor  : AMD EPYC 7R32
    CPU cores  : 2 @ 2799.990 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ❌ Disabled
    RAM        : 3.8 GiB
    Swap       : 0.0 KiB
    Disk       : 96.9 GiB
    
    fio Disk Speed Tests (Mixed R/W 50/50):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 10.31 MB/s    (2.5k) | 164.19 MB/s   (2.5k)
    Write      | 10.34 MB/s    (2.5k) | 165.05 MB/s   (2.5k)
    Total      | 20.66 MB/s    (5.1k) | 329.24 MB/s   (5.1k)
               |                      |
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 207.90 MB/s    (406) | 193.55 MB/s    (189)
    Write      | 218.95 MB/s    (427) | 206.44 MB/s    (201)
    Total      | 426.85 MB/s    (833) | 399.99 MB/s    (390)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed
                    |                           |                 |
    Clouvider       | London, UK (10G)          | 713 Mbits/sec   | 383 Mbits/sec
    Online.net      | Paris, FR (10G)           | 1.28 Gbits/sec  | 589 Mbits/sec
    Clouvider       | NYC, NY, US (10G)         | 436 Mbits/sec   | 376 Mbits/sec
    
    Geekbench 5 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 989
    Multi Core      | 1251
    Full Test       | https://browser.geekbench.com/v5/cpu/13601044
    


    Note:
    • EC2 price does not include the disk
    • EC2 instance use io1 EBS type with 5000 IOPS
  • @JeDaYoshi said:

    I still don't understand why the clock speed is so low, though.

    Epyc 7713 base clock is 2.0 GHz, but can be boosted to 3.675 GHz. YABS only show the base clock I think.

  • edited March 2022

    @kuroneko23 said:

    @JeDaYoshi said:

    I still don't understand why the clock speed is so low, though.

    Epyc 7713 base clock is 2.0 GHz, but can be boosted to 3.675 GHz. YABS only show the base clock I think.

    It shows the current CPU clock speed. Generally you'd have expected them to set it to always 3.675 GHz, like most providers having a CPU that can boost frequencies.

    The other YABS looks promising, but I'm still not fond of the disk speed.

  • Boost freq is only available when temperature allows it. If you run something compute intensive on all cores, the chip heats up and runs at base freq. Threadripper is single socket but has higher base freq so they are the fastest single chips.

Sign In or Register to comment.