SPECjbb MultiJVM - Java Performance

Moving on from SPECCPU, we shift over to SPECjbb2015. SPECjbb is a from ground-up developed benchmark that aims to cover both Java performance and server-like workloads, from the SPEC website:

“The SPECjbb2015 benchmark is based on the usage model of a worldwide supermarket company with an IT infrastructure that handles a mix of point-of-sale requests, online purchases, and data-mining operations. It exercises Java 7 and higher features, using the latest data formats (XML), communication using compression, and secure messaging.

Performance metrics are provided for both pure throughput and critical throughput under service-level agreements (SLAs), with response times ranging from 10 to 100 milliseconds.”

The important thing to note here is that the workload is of a transactional nature that mostly works on the data-plane, between different Java virtual machines, and thus threads.

We’re using the MultiJVM test method where as all the benchmark components, meaning controller, server and client virtual machines are running on the same physical machine.

The JVM runtime we’re using is OpenJDK 15 on both x86 and Arm platforms, although not exactly the same sub-version, but closest we could get:

EPYC & Xeon systems:

openjdk 15 2020-09-15
OpenJDK Runtime Environment (build 15+36-Ubuntu-1)
OpenJDK 64-Bit Server VM (build 15+36-Ubuntu-1, mixed mode, sharing)

Altra system:

openjdk 15.0.1 2020-10-20
OpenJDK Runtime Environment 20.9 (build 15.0.1+9)
OpenJDK 64-Bit Server VM 20.9 (build 15.0.1+9, mixed mode, sharing)

Furthermore, we’re configuring SPECjbb’s runtime settings with the following configurables:

SPEC_OPTS_C="-Dspecjbb.group.count=$GROUP_COUNT -Dspecjbb.txi.pergroup.count=$TI_JVM_COUNT -Dspecjbb.forkjoin.workers=N -Dspecjbb.forkjoin.workers.Tier1=N -Dspecjbb.forkjoin.workers.Tier2=1 -Dspecjbb.forkjoin.workers.Tier3=16"

Where N=160 for 2S Altra test runs, N=80 for 1S Altra test runs, N=112 for 2S Xeon 8280, N=56 for 1S Xeon 8280, and N=128 for 2S and 1S on the EPYC system. The 75F3 system had the worker count reduced to 64 and 32 for 2S/1S runs.

The Xeon 8380 was running at N=140 for 2S Xeon 8380 and N=70 for 1S - the benchmark had been erroring out at higher thread counts.

In terms of JVM options, we’re limiting ourselves to bare-bone options to keep things simple and straightforward:

EPYC & Altra systems:

JAVA_OPTS_C="-server -Xms2g -Xmx2g -Xmn1536m -XX:+UseParallelGC "
JAVA_OPTS_TI="-server -Xms2g -Xmx2g -Xmn1536m -XX:+UseParallelGC"
JAVA_OPTS_BE="-server -Xms48g -Xmx48g -Xmn42g -XX:+UseParallelGC -XX:+AlwaysPreTouch"

Xeon Cascade Lake systems:

JAVA_OPTS_C="-server -Xms2g -Xmx2g -Xmn1536m -XX:+UseParallelGC"
JAVA_OPTS_TI="-server -Xms2g -Xmx2g -Xmn1536m -XX:+UseParallelGC"
JAVA_OPTS_BE="-server -Xms172g -Xmx172g -Xmn156g -XX:+UseParallelGC -XX:+AlwaysPreTouch"

Xeon Ice Lake systems (SNC1):

JAVA_OPTS_C="-server -Xms2g -Xmx2g -Xmn1536m -XX:+UseParallelGC"
JAVA_OPTS_TI="-server -Xms2g -Xmx2g -Xmn1536m -XX:+UseParallelGC"
JAVA_OPTS_BE="-server -Xms192g -Xmx192g -Xmn168g -XX:+UseParallelGC -XX:+AlwaysPreTouch"

Xeon Ice Lake systems (SNC2):

JAVA_OPTS_C="-server -Xms2g -Xmx2g -Xmn1536m -XX:+UseParallelGC"
JAVA_OPTS_TI="-server -Xms2g -Xmx2g -Xmn1536m -XX:+UseParallelGC"
JAVA_OPTS_BE="-server -Xms96g -Xmx96g -Xmn84g -XX:+UseParallelGC -XX:+AlwaysPreTouch"

The reason the Xeon CLX system is running a larger back-end heap is because we’re running a single NUMA node per socket, while for the Altra and EPYC we’re running four NUMA nodes per socket for maximised throughput, meaning for the 2S figures we have 8 backends running for the Altra and EPYC and 2 for the Xeon, and naturally half of those numbers for the 1S benchmarks.

For the Ice Lake system, I ran both SNC1 (one NUMA node) as SNC2 (two nodes), with the corresponding scaling in the back-end memory allocation.

The back-ends and transaction injectors are affinitised to their local NUMA node with numactl –cpunodebind and –membind, while the controller is called with –interleave=all.

The max-jOPS and critical-jOPS result figures are defined as follows:

"The max-jOPS is the last successful injection rate before the first failing injection rate where the reattempt also fails. For example, if during the RT-curve phase the injection rate of 80000 passes, but the next injection rate of 90000 fails on two successive attempts, then the max-jOPS would be 80000."

"The overall critical-jOPS is computed by taking the geomean of the individual critical-jOPS computed at these five SLA points, namely:

      • Critical-jOPSoverall = Geo-mean of (critical-jOPS@ 10ms, 25ms, 50ms, 75ms and 100ms response time SLAs)

During the RT curve building phase the Transaction Injector measures the 99th percentile response times at each step level for all the requests (see section 9) that are considered in the metrics computations. It then computes the Critical-jOPS for each of the above five SLA points using the following formula:
(first * nOver + last * nUnder) / (nOver + nUnder) "


That’s a lot of technicalities to explain an admittedly complex benchmark, but the gist of it is that max-jOPS represents the maximum transaction throughput of a system until further requests fail, and critical-jOPS is an aggregate geomean transaction throughput within several levels of guaranteed response times, essentially different levels of quality of service.

Beyond the result figures, the benchmark keeps detailed track of timings of responses and tracks a few important statistical data-points across a response-time curve, as follows:


2S Xeon 8380 THP Enabled


2S Xeon 8280 THP Enabled 

Comparing the Xeon 8380 to the Xeon 8280, what’s to be immediately noted is the much-improved maximum throughput figure of the new part, scaling at +64% compared to its predecessor. We’re seeing that the load slope where the 99th percentile SLA figures rises comes in at a relative earlier point, and the corresponding critical-jOPS point lands in relatively earlier than the Xeon 8280.


2S EPYC 7763 THP Enabled


2S Altra Q80-33 THP Enabled

I included the AMD EPYC 7763 and Altra graphs for context.

SPECjbb2015-MultiJVM max-jOPS

 As commented in the response curve analysis, the new Xeon 8380 sees huge leaps in the max-jOPS metrics, vastly outperforming the Xeon 8280 and landing in a very favourable competitive positioning compared to the AMD parts.

My theory here is that because of the good per-core performance of the Intel design, along with the monolithic mesh architecture, while Intel doesn’t quite catch up with AMD, it performs very well with relatively significantly fewer cores.

SPECjbb2015-MultiJVM critical-jOPS

 The critical-jOPS metric however wasn’t quite as positive for the new Xeon 8380. Although the chip is showing increases in performance, there’s not as strong as the max-jOPS measurements. At first I had measured the SNC off mode of the platform, similar to the 8280 numbers we have (Our ASRock test bed doesn’t expose SNC options in the BIOS), however these results were extremely meagre as they barely differentiated to the 8280. Running the system in SNC2 mode actually improved the critical jOPS figure more significantly, whilst only marginally affecting the max-jOPS metric.

What’s really odd about the results though is that this larger increase only happens in the 2S test figures, with the 1S being unfavourable to the new Ice Lake part, losing to the 8280 in both modes. I had repeated these numbers several times to be sure they’re repeatable, and they were indeed so – as odd as that is. The 1S reduction in the critical-jOPS could be explained through the larger mesh size and larger core count of the 8380, and we did see slight regressions in core-to-core latencies. If the mesh intersection bandwidth did not increase with its size, that also could be a culprit of these figures, as the workload is hammering core-to-core transactions as well as the L3 cache of the chip.

Why the 2S figures see a bigger advantage of migrating to SNC2 could be a result of how on-chip traffic is routed, as well as the traffic flows through the UPI link blocks of the chip – at least that would be my working hypothesis.

Intel had disclosed a +62% figure for a “Java Throughput under SLA” workload they wouldn’t specify, and this does track well with our max-jOPS results. While the critical-jOPS increases seem a bit disappointing generationally, how it translates to the real world in contrast to the max-jOPS figure depends on how strict one’s SLA metrics are.

SPEC - Per-Core Performance under Load Compiling LLVM, NAMD Performance
Comments Locked

169 Comments

View All Comments

  • ricebunny - Tuesday, April 6, 2021 - link

    See it like it this: the benchmark is a racing track, the CPU is a car and the compiler is the driver. If I want to get the best time for each car on a given track I will not have them driven by the same driver. Rather, I will get the best driver for each car. A single driver will repeat the same mistakes in both cars, but one car may be more forgiving than the other.
  • DigitalFreak - Tuesday, April 6, 2021 - link

    Is the compiler called The Stig?
  • Wilco1 - Tuesday, April 6, 2021 - link

    Then you are comparing drivers and not cars. A good driver can win a race with a slightly slower car. And I know a much faster driver that can beat your best driver. And he will win even with a much slower car. So does the car really matter as long as you have a really good driver?

    In the real world we compare cars by subjecting them to identical standardized tests rather than having a grandma drive one car and Lewis Hamilton drive another when comparing their performance/efficiency/acceleration/safety etc.
  • Makste - Wednesday, April 7, 2021 - link

    Well said
  • ricebunny - Wednesday, April 7, 2021 - link

    Based on the compiler options that Anandtech used, we already have the situation that Intel and AMD CPUs are executing different code for the same benchmark. From there it’s only a small step further to use the best compiler for each CPU.
  • mode_13h - Wednesday, April 7, 2021 - link

    So, you're saying make the situation MORE lopsided? Instead, maybe they SHOULD use the same compiled code!
  • mode_13h - Wednesday, April 7, 2021 - link

    This is a dumb analogy. CPUs are not like race cars. They're more like family sedans or maybe 18-wheeler semi trucks (in the case of server CPUs). As such, they should be tested the way most people are going to use them.

    And almost NOBODY is compiling all their software with ICC. I almost never even hear about ICC, any more.

    I'm even working with an Intel applications engineer on a CPU performance problem, and even HE doesn't tell me to build their own Intel-developed software with ICC!
  • KurtL - Wednesday, April 7, 2021 - link

    Using identical compilers is the most unfair option there is to compare CPUs. Hardware and software on a modern system is tightly connected so it only makes sense to use those compilers on each platform that also are best optimised for that particular platform. Using a compiler that is underdeveloped for one platform is what makes an unfair comparison.
  • Makste - Wednesday, April 7, 2021 - link

    I think that using one unoptimized compiler for both is the best way to judge their performance. Such a compiler rules out bias and concentrates on pure hardware capabilities
  • ricebunny - Wednesday, April 7, 2021 - link

    You do realize that even the same gcc compiler with the settings that Anandtech used will generate different machine code for Intel and AMD architectures, let alone for ARM? To really make it "apples-to-apples" on Linux x86 they should've used "--with-tune=generic" option: then both CPUs will execute the exact same code.

    But personally, I would prefer that they generated several binaries for each test, built them with optimal settings for each of the commonly used compilers: gcc, icc, aocc on Linux and perhaps even msvc on Windows. It's a lot more work I know, but I would appreciate it :)

Log in

Don't have an account? Sign up now